GPT & Haskell

newer
IFL23 first CFP

older
Re: [Haskell-cafe] GPT & Haskell

Gregory Guthrie

31 Mar 2023 31 Mar '23

8:02 p.m.

...

From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable.

Still... -------------------------------------------------------------------------------------------------- GPT-4 when prompted: Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states. gives the following response: -- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element -- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned The above code runs: https://onecompiler.com/cpp/3z48yt2pp [cid:image001.png@01D963E1.B0EB6AB0]

Attachments:

attachment.html (text/html — 4.5 KB)
image001.png (image/png — 38.2 KB)

Show replies by date

Jeffrey Brown

31 Mar 31 Mar

8:33 p.m.

For toy problems they can impress, but getting an LLM to understand a realistic codebase well enough to generate useful changes seems impossible. On Fri, Mar 31, 2023 at 3:04 PM Gregory Guthrie wrote:

...

From a colleague:

GPT Pretty smart(?!).

But this is of course simple, and searchable.

Still…

--------------------------------------------------------------------------------------------------

GPT-4 when prompted:

*Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.*

gives the following response:

*-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element -- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned*

The above code runs:

*https://onecompiler.com/cpp/3z48yt2pp https://onecompiler.com/cpp/3z48yt2pp*

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Jeff Brown | Jeffrey Benjamin Brown LinkedIn https://www.linkedin.com/in/jeffreybenjaminbrown | Github https://github.com/jeffreybenjaminbrown | Twitter https://twitter.com/carelogic | Facebook https://www.facebook.com/mejeff.younotjeff

Brandon Allbery

8:40 p.m.

We tried to get it to rewrite XMonad.Layout.PerScreen to properly work per screen instead of the `ifWidth` hack. (It can use the passed-in screen rectangle to do this.) It didn't even understand the question. On Fri, Mar 31, 2023 at 4:34 PM Jeffrey Brown wrote:

...

For toy problems they can impress, but getting an LLM to understand a realistic codebase well enough to generate useful changes seems impossible.

On Fri, Mar 31, 2023 at 3:04 PM Gregory Guthrie wrote:

...
From a colleague:

GPT Pretty smart(?!).

But this is of course simple, and searchable.

Still…

--------------------------------------------------------------------------------------------------

GPT-4 when prompted:

*Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.*

gives the following response:

*-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element -- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned*

The above code runs:

*https://onecompiler.com/cpp/3z48yt2pp https://onecompiler.com/cpp/3z48yt2pp*

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Jeff Brown | Jeffrey Benjamin Brown LinkedIn https://www.linkedin.com/in/jeffreybenjaminbrown | Github https://github.com/jeffreybenjaminbrown | Twitter https://twitter.com/carelogic | Facebook https://www.facebook.com/mejeff.younotjeff _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- brandon s allbery kf8nh allbery.b@gmail.com

William Yager

9:17 p.m.

I found it to be quite good at questions without too large of a question-specific context. E.g. it can easily solve questions like "Write a `Functor` instance for `data Foo a = Bar a | Baz a a` and prove it obeys the functor laws.". No doubt the models will get better at handling larger problems as they are scaled up.

Gregory Guthrie

9:15 p.m.

Yep! Programmers still have jobs. Designers even more. Dr. Gregory Guthrie From: Jeffrey Brown Sent: Friday, March 31, 2023 3:34 PM To: Gregory Guthrie Cc: haskell-cafe@haskell.org; Nenad Končar ; Haskell Community Subject: Re: [Haskell-cafe] GPT & Haskell For toy problems they can impress, but getting an LLM to understand a realistic codebase well enough to generate useful changes seems impossible. On Fri, Mar 31, 2023 at 3:04 PM Gregory Guthrie mailto:guthrie@miu.edu> wrote: From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable. Still… -------------------------------------------------------------------------------------------------- GPT-4 when prompted: Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states. gives the following response: -- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element -- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned The above code runs: https://onecompiler.com/cpp/3z48yt2pp [cid:image001.png@01D963E7.7BCC1DB0] _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post. -- Jeff Brown | Jeffrey Benjamin Brown LinkedInhttps://www.linkedin.com/in/jeffreybenjaminbrown | Githubhttps://github.com/jeffreybenjaminbrown | Twitterhttps://twitter.com/carelogic | Facebookhttps://www.facebook.com/mejeff.younotjeff

Levent Erkok

9:24 p.m.

This code is buggy when lo + hi overflows your integer type. Classic (extremely) corner case bug: https://ai.googleblog.com/2006/06/extra-extra-read-all-about-it-nearly.html

...

On Mar 31, 2023, at 1:02 PM, Gregory Guthrie wrote:

From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable.

Still… -------------------------------------------------------------------------------------------------- GPT-4 when prompted:

Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.

gives the following response:

-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space

binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element

-- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned

The above code runs:

https://onecompiler.com/cpp/3z48yt2pp

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Yasuaki Kudo

9:37 p.m.

I am really curious - is it possible to create a programming language designed from the ground up to be used by GPT and similar tools? - Easy to specify the desired outcome - Easy to verify results I tried GPT4 to produce COQ proofs of Yoneda Lemma but it could not even complete the interfaces (without implementation) - they failed to compile, by being tripped up by implicit parameters etc. But GPT4 gets much closer to completion compared to GPT3.5 I wonder what "GPT-native" programming language might look like😄 -Yasu

...

On Apr 1, 2023, at 06:24, Levent Erkok wrote:

This code is buggy when lo + hi overflows your integer type. Classic (extremely) corner case bug: https://ai.googleblog.com/2006/06/extra-extra-read-all-about-it-nearly.html

...
On Mar 31, 2023, at 1:02 PM, Gregory Guthrie wrote:

From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable.

Still… -------------------------------------------------------------------------------------------------- GPT-4 when prompted:

Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.

gives the following response:

-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space

binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element

-- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned

The above code runs:

https://onecompiler.com/cpp/3z48yt2pp

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Branimir Maksimovic

9:42 p.m.

I was thoughtt from my mathematic logic proffesor in 1987 that algorithm for making algorithms doesn’t exist. He proved that there is no algorihm for provin hieor order logic formulas… Greets, Branmir.

...

On 31. 3. 2023., at 23:37, Yasuaki Kudo wrote:

I am really curious - is it possible to create a programming language designed from the ground up to be used by GPT and similar tools?

- Easy to specify the desired outcome - Easy to verify results

I tried GPT4 to produce COQ proofs of Yoneda Lemma but it could not even complete the interfaces (without implementation) - they failed to compile, by being tripped up by implicit parameters etc.

But GPT4 gets much closer to completion compared to GPT3.5

I wonder what "GPT-native" programming language might look like😄

-Yasu

...
On Apr 1, 2023, at 06:24, Levent Erkok wrote:

This code is buggy when lo + hi overflows your integer type. Classic (extremely) corner case bug: https://ai.googleblog.com/2006/06/extra-extra-read-all-about-it-nearly.html

...
On Mar 31, 2023, at 1:02 PM, Gregory Guthrie wrote:

From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable.

Still… -------------------------------------------------------------------------------------------------- GPT-4 when prompted:

Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.

gives the following response:

-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space

binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element

-- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned

The above code runs:

https://onecompiler.com/cpp/3z48yt2pp

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Alexey Vagarenko

11:01 p.m.

Isn’t binary search useless for lists? It performs worse than full list traversal. Сб, 1 апр. 2023 г. в 03:43, Branimir Maksimovic < branimir.maksimovic@gmail.com>:

...

I was thoughtt from my mathematic logic proffesor in 1987 that algorithm for making algorithms doesn’t exist. He proved that there is no algorihm for provin hieor order logic formulas…

Greets, Branmir.

On 31. 3. 2023., at 23:37, Yasuaki Kudo wrote:

I am really curious - is it possible to create a programming language designed from the ground up to be used by GPT and similar tools?

- Easy to specify the desired outcome - Easy to verify results

I tried GPT4 to produce COQ proofs of Yoneda Lemma but it could not even complete the interfaces (without implementation) - they failed to compile, by being tripped up by implicit parameters etc.

But GPT4 gets much closer to completion compared to GPT3.5

I wonder what "GPT-native" programming language might look like😄

-Yasu

On Apr 1, 2023, at 06:24, Levent Erkok wrote:

This code is buggy when lo + hi overflows your integer type. Classic (extremely) corner case bug: https://ai.googleblog.com/2006/06/extra-extra-read-all-about-it-nearly.html

On Mar 31, 2023, at 1:02 PM, Gregory Guthrie wrote:

From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable.

Still…

--------------------------------------------------------------------------------------------------

GPT-4 when prompted: *Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.*

gives the following response:

*-- Binary search in Haskell-- Precondition: the list is sorted in ascending order-- Postcondition: the function returns the index of the target element if found, or Nothing otherwise-- Invariant: the search space is always within the bounds of the list-- Variant: the search space decreases by half at each recursive call-- Internal state: the lower and upper bounds of the search spacebinarySearch :: Ord a => [a] -> a -> Maybe IntbinarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element-- Print results using putStrLnmain :: IO ()main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned*

The above code runs: *https://onecompiler.com/cpp/3z48yt2pp https://onecompiler.com/cpp/3z48yt2pp*

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Branimir Maksimovic

11:07 p.m.

Of course it’ useless. GPT should be tested with something that can’t be found on internet :P I tried with problem euler 512, it can’t find optimal solution, but can perform microoptimisations of existing solution, which is still impressive :P Greets, Branimir.

...

On 1. 4. 2023., at 01:01, Alexey Vagarenko wrote:

Isn’t binary search useless for lists? It performs worse than full list traversal.

Сб, 1 апр. 2023 г. в 03:43, Branimir Maksimovic mailto:branimir.maksimovic@gmail.com>:

...
I was thoughtt from my mathematic logic proffesor in 1987 that algorithm for making algorithms doesn’t exist. He proved that there is no algorihm for provin hieor order logic formulas…

Greets, Branmir.

...
On 31. 3. 2023., at 23:37, Yasuaki Kudo mailto:yasu@yasuaki.com> wrote:

I am really curious - is it possible to create a programming language designed from the ground up to be used by GPT and similar tools?

- Easy to specify the desired outcome - Easy to verify results

I tried GPT4 to produce COQ proofs of Yoneda Lemma but it could not even complete the interfaces (without implementation) - they failed to compile, by being tripped up by implicit parameters etc.

But GPT4 gets much closer to completion compared to GPT3.5

I wonder what "GPT-native" programming language might look like😄

-Yasu

...
On Apr 1, 2023, at 06:24, Levent Erkok mailto:erkokl@gmail.com> wrote:

This code is buggy when lo + hi overflows your integer type. Classic (extremely) corner case bug: https://ai.googleblog.com/2006/06/extra-extra-read-all-about-it-nearly.html

...
On Mar 31, 2023, at 1:02 PM, Gregory Guthrie mailto:guthrie@miu.edu> wrote:

From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable.

Still… -------------------------------------------------------------------------------------------------- GPT-4 when prompted:

Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.

gives the following response:

-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space

binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element

-- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned

The above code runs:

https://onecompiler.com/cpp/3z48yt2pp

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Yasuaki Kudo

11:24 p.m.

Mig Mit

11:27 p.m.

Jeff Clites

11:56 p.m.

MigMit

1 Apr 1 Apr

12:08 a.m.

That's true. But I feel like a lot of progress in programming was made in attempt to make "non-technical people" communicate with a computer without a programmer in between. So far, despite all that was done, those attempts did not really succeed. I'm sceptical regarding AI chances, honestly. I mean, just imagine a business person suddenly finding out they can't improve the computer system by yelling at the programmer.

...

On 1 Apr 2023, at 01:56, Jeff Clites wrote:

Yes, if you’re a programmer. A lot of real-world business scenarios are basically a non-technical person using English to give a (flawed, incomplete) specification, and a programmer using their expertise not only to implement but also to fill in the incomplete parts of the specification and also to determine what missing parts of the specification can’t be interpolated and have to be asked about. That sort of filling in missing parts based on precedent is the exact sort of thing these AI are good at.

Jeff

...
On Mar 31, 2023, at 4:27 PM, Mig Mit wrote:

How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms. It is way easier to just code the damn thing.

Sent from my iPad

...
On 2023. Apr 1., at 1:24, Yasuaki Kudo wrote:

There is TTS (text to speech) and STT (speech to text) Former is easier, latter is harder.

Similarly, is it possible that we generate a huge variations of English explanation of correct compiled programs (like TTS - easy-ish, because we can program English explanations algorithmically ) and the get GPT-like-things to learn from that huge sample so it can convert human instructions to accurate programs? (like STT - hard),

...
On Apr 1, 2023, at 08:08, Branimir Maksimovic wrote:

Of course it’ useless. GPT should be tested with something that can’t be found on internet :P

I tried with problem euler 512, it can’t find optimal solution, but can perform microoptimisations of existing solution, which is still impressive :P

Greets, Branimir.

...
On 1. 4. 2023., at 01:01, Alexey Vagarenko wrote:

Isn’t binary search useless for lists? It performs worse than full list traversal.

Сб, 1 апр. 2023 г. в 03:43, Branimir Maksimovic : I was thoughtt from my mathematic logic proffesor in 1987 that algorithm for making algorithms doesn’t exist. He proved that there is no algorihm for provin hieor order logic formulas…

Greets, Branmir.

...
On 31. 3. 2023., at 23:37, Yasuaki Kudo wrote:

I am really curious - is it possible to create a programming language designed from the ground up to be used by GPT and similar tools?

- Easy to specify the desired outcome - Easy to verify results

I tried GPT4 to produce COQ proofs of Yoneda Lemma but it could not even complete the interfaces (without implementation) - they failed to compile, by being tripped up by implicit parameters etc.

But GPT4 gets much closer to completion compared to GPT3.5

I wonder what "GPT-native" programming language might look like😄

-Yasu

> On Apr 1, 2023, at 06:24, Levent Erkok wrote: > > This code is buggy when lo + hi overflows your integer type. Classic (extremely) corner case bug: https://ai.googleblog.com/2006/06/extra-extra-read-all-about-it-nearly.html > >> On Mar 31, 2023, at 1:02 PM, Gregory Guthrie wrote: >> >> From a colleague: >> GPT Pretty smart(?!). >> But this is of course simple, and searchable. >> Still… >> -------------------------------------------------------------------------------------------------- >> GPT-4 when prompted: >> Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states. >> gives the following response: >> -- Binary search in Haskell >> -- Precondition: the list is sorted in ascending order >> -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise >> -- Invariant: the search space is always within the bounds of the list >> -- Variant: the search space decreases by half at each recursive call >> -- Internal state: the lower and upper bounds of the search space >> >> binarySearch :: Ord a => [a] -> a -> Maybe Int >> binarySearch xs x = go xs x 0 (length xs - 1) >> where >> go xs x lo hi >> | lo > hi = Nothing -- base case: search space is empty >> | x == midElem = Just mid -- base case: target element is found >> | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half >> | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half >> where >> mid = (lo + hi) `div` 2 -- calculate the middle index >> midElem = xs !! mid -- get the middle element >> >> -- Print results using putStrLn >> main :: IO () >> main = do >> let xs = [1, 3, 5, 7, 9] -- a sorted list of integers >> let x = 5 -- a target element to search for >> let result = binarySearch xs x -- call the binary search function >> case result of -- pattern match on the result >> Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned >> Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned >> The above code runs: >> https://onecompiler.com/cpp/3z48yt2pp >> >> _______________________________________________ >> Haskell-Cafe mailing list >> To (un)subscribe, modify options or view archives go to: >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe >> Only members subscribed via the mailman list are allowed to post. > > > _______________________________________________ > Haskell-Cafe mailing list > To (un)subscribe, modify options or view archives go to: > http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe > Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post. _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

William Yager

3:36 a.m.

On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote:

...

How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...

I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try. I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug. Very useful!

Aloïs Cochard

6:18 a.m.

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends. I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf Be ready for a good laugh Sincerely yours On Sat, 1 Apr 2023 at 05:36, William Yager wrote:

...

On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote:

...
How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

Dominik Schrempf

9:44 a.m.

Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me! Dominik On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" wrote:

...

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Sincerely yours

On Sat, 1 Apr 2023 at 05:36, William Yager wrote:

...
On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote:

...
How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

Aloïs Cochard

11:36 a.m.

I don't know, personally I could not stop laughing at the answer from version 3.5... it's so freaking stupid! But if like Dominik you are scared after reading that part you might be interested to look into: https://futureoflife.org/open-letter/pause-giant-ai-experiments/ On Sat, 1 Apr 2023 at 11:44, Dominik Schrempf wrote:

...

Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me!

Dominik

On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" < alois.cochard@gmail.com> wrote:

...
How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Sincerely yours

On Sat, 1 Apr 2023 at 05:36, William Yager wrote:

...
On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote:

...
How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

MigMit

11:51 a.m.

Or this: https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/ (to which somebody said "I work in a datacenter. So far no AI threatened violence on me; Yudkowsky just did")

...

On 1 Apr 2023, at 13:36, Aloïs Cochard wrote:

I don't know, personally I could not stop laughing at the answer from version 3.5... it's so freaking stupid!

But if like Dominik you are scared after reading that part you might be interested to look into: https://futureoflife.org/open-letter/pause-giant-ai-experiments/

On Sat, 1 Apr 2023 at 11:44, Dominik Schrempf wrote: Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me!

Dominik

On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" wrote: How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Sincerely yours

On Sat, 1 Apr 2023 at 05:36, William Yager wrote: On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote: How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

Tom Ellis

12:24 p.m.

I never thought I would see Time publish an article advocating that "allied nuclear countries [should be] willing to run some risk of nuclear exchange if that’s what it takes to reduce the risk of large AI training runs", written by a man who didn't go to high school. Tom On Sat, Apr 01, 2023 at 01:51:54PM +0200, MigMit wrote:

...

Or this: https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/

(to which somebody said "I work in a datacenter. So far no AI threatened violence on me; Yudkowsky just did")

...
On 1 Apr 2023, at 13:36, Aloïs Cochard wrote:

I don't know, personally I could not stop laughing at the answer from version 3.5... it's so freaking stupid!

But if like Dominik you are scared after reading that part you might be interested to look into: https://futureoflife.org/open-letter/pause-giant-ai-experiments/

On Sat, 1 Apr 2023 at 11:44, Dominik Schrempf wrote: Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me!

Dominik

On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" wrote: How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Aloïs Cochard

12:51 p.m.

Very interesting, thanks for sharing. This is a very radical viewpoint, I think there is tons of research that can be done on alignment and for this reason I don't think we should shut it all down. Obviously it's not a good idea to give access to such a system to everyone, I hope that at least we all start to understand this. On Sat, 1 Apr 2023 at 13:52, MigMit wrote:

...

Or this: https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/

(to which somebody said "I work in a datacenter. So far no AI threatened violence on me; Yudkowsky just did")

...
On 1 Apr 2023, at 13:36, Aloïs Cochard wrote:

I don't know, personally I could not stop laughing at the answer from version 3.5... it's so freaking stupid!

But if like Dominik you are scared after reading that part you might be interested to look into: https://futureoflife.org/open-letter/pause-giant-ai-experiments/

On Sat, 1 Apr 2023 at 11:44, Dominik Schrempf < dominik.schrempf@gmail.com> wrote: Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me!

Dominik

On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" < alois.cochard@gmail.com> wrote: How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Sincerely yours

On Sat, 1 Apr 2023 at 05:36, William Yager wrote: On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote: How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

MigMit

12:57 p.m.

Why is it obvious?

...

On 1 Apr 2023, at 14:51, Aloïs Cochard wrote:

Very interesting, thanks for sharing.

This is a very radical viewpoint, I think there is tons of research that can be done on alignment and for this reason I don't think we should shut it all down.

Obviously it's not a good idea to give access to such a system to everyone, I hope that at least we all start to understand this.

On Sat, 1 Apr 2023 at 13:52, MigMit wrote: Or this: https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/

(to which somebody said "I work in a datacenter. So far no AI threatened violence on me; Yudkowsky just did")

...
On 1 Apr 2023, at 13:36, Aloïs Cochard wrote:

I don't know, personally I could not stop laughing at the answer from version 3.5... it's so freaking stupid!

But if like Dominik you are scared after reading that part you might be interested to look into: https://futureoflife.org/open-letter/pause-giant-ai-experiments/

On Sat, 1 Apr 2023 at 11:44, Dominik Schrempf wrote: Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me!

Dominik

On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" wrote: How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Sincerely yours

On Sat, 1 Apr 2023 at 05:36, William Yager wrote: On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote: How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

Aloïs Cochard

1:08 p.m.

...

From a theoretical point of view, this can be observed in https://arxiv.org/pdf/2303.12712.pdf from page 126 to the end of the paper.

More concretely the scale at which such a system can be used to synthesis information or "predict" stock market is alarming, those are just some random examples. On Sat, 1 Apr 2023 at 14:57, MigMit wrote:

...

Why is it obvious?

...
On 1 Apr 2023, at 14:51, Aloïs Cochard wrote:

Very interesting, thanks for sharing.

This is a very radical viewpoint, I think there is tons of research that can be done on alignment and for this reason I don't think we should shut it all down.

Obviously it's not a good idea to give access to such a system to everyone, I hope that at least we all start to understand this.

On Sat, 1 Apr 2023 at 13:52, MigMit wrote: Or this: https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/

(to which somebody said "I work in a datacenter. So far no AI threatened violence on me; Yudkowsky just did")

...
On 1 Apr 2023, at 13:36, Aloïs Cochard wrote:

I don't know, personally I could not stop laughing at the answer from version 3.5... it's so freaking stupid!

But if like Dominik you are scared after reading that part you might be interested to look into: https://futureoflife.org/open-letter/pause-giant-ai-experiments/

On Sat, 1 Apr 2023 at 11:44, Dominik Schrempf < dominik.schrempf@gmail.com> wrote: Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me!

Dominik

On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" < alois.cochard@gmail.com> wrote: How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Sincerely yours

On Sat, 1 Apr 2023 at 05:36, William Yager wrote: On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote: How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

MigMit

1:32 p.m.

Um, I see someone who can do some math. No disrespect, but I think I can do better. Should I also be under this "restricted access" policy? And using software to predict stock market is done constantly; software does it better than any human (and probably better than any AI), so, what's the problem here?

...

On 1 Apr 2023, at 15:08, Aloïs Cochard wrote:

From a theoretical point of view, this can be observed in https://arxiv.org/pdf/2303.12712.pdf from page 126 to the end of the paper.

More concretely the scale at which such a system can be used to synthesis information or "predict" stock market is alarming, those are just some random examples.

On Sat, 1 Apr 2023 at 14:57, MigMit wrote: Why is it obvious?

...
On 1 Apr 2023, at 14:51, Aloïs Cochard wrote:

Very interesting, thanks for sharing.

This is a very radical viewpoint, I think there is tons of research that can be done on alignment and for this reason I don't think we should shut it all down.

Obviously it's not a good idea to give access to such a system to everyone, I hope that at least we all start to understand this.

On Sat, 1 Apr 2023 at 13:52, MigMit wrote: Or this: https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/

(to which somebody said "I work in a datacenter. So far no AI threatened violence on me; Yudkowsky just did")

...
On 1 Apr 2023, at 13:36, Aloïs Cochard wrote:

I don't know, personally I could not stop laughing at the answer from version 3.5... it's so freaking stupid!

But if like Dominik you are scared after reading that part you might be interested to look into: https://futureoflife.org/open-letter/pause-giant-ai-experiments/

On Sat, 1 Apr 2023 at 11:44, Dominik Schrempf wrote: Well, it seems like GPT4 gets most of the questions right. This was more of a scary read than a funny one to me!

Dominik

On April 1, 2023 8:18:08 AM GMT+02:00, "Aloïs Cochard" wrote: How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Sincerely yours

On Sat, 1 Apr 2023 at 05:36, William Yager wrote: On Fri, Mar 31, 2023 at 7:27 PM Mig Mit wrote: How is it going to benefit anyone though? English, like all natural languages, is terrible for describing algorithms.

The other day, I asked ChatGPT to solve a boring but non-trivial programming problem I had:

...
I have a png file which contains a bunch of 5x7 bitmap font characters. Write a python program which reads in the png, and then for each character in the png, extracts the 35 bits of each character and assembles them into an integer. The final output should be a list of integers, each integer containing a binary representation of each character in the font.

It did this basically perfectly. It wasn't precisely matched to my format, but only because I was ambiguous in the problem statement. I gave it a little more context about the image format and it did it exactly right on the second try.

I also asked it to write an inverse function of this, which it did perfectly in Python, and did pretty well in Rust, although with a small and understandable bug.

Very useful! _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

-- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

Will Yager

1:48 p.m.

...

On Apr 1, 2023, at 02:18, Aloïs Cochard wrote:

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw???

The same question applies to hiring human programmers.

...

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt:

I'm not sure why you classify successfully getting practical tasks done faster as "wasting time". Best, Will

MigMit

1:57 p.m.

Well, human programmers can be shamed, yelled at, fired (which would actually hurt them), or, in extreme cases, prosecuted. They have every insentive to do their job right. It's not a huge difference, but still.

...

On 1 Apr 2023, at 15:48, Will Yager wrote:

...
On Apr 1, 2023, at 02:18, Aloïs Cochard wrote:

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw???

The same question applies to hiring human programmers.

...
I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt:

I'm not sure why you classify successfully getting practical tasks done faster as "wasting time".

Best, Will

Will Yager

2:09 p.m.

...

On Apr 1, 2023, at 09:57, MigMit wrote:

Well, human programmers can be shamed, yelled at, fired (which would actually hurt them), or, in extreme cases, prosecuted. They have every insentive to do their job right.

ChatGPT has RLHF. It has incentives to do its job right as well.

MigMit

2:16 p.m.

Depends. In the current arrangement, people who do RLHF (by the way, that's the first time I've seen this abbreviation, I'm sure) aren't likely to know it did a bad job. If it's something under your control, yes, I would agree. Although we now headed straight into the AI rights territory (a programmer who is being yelled at a little too much can just quit).

...

On 1 Apr 2023, at 16:09, Will Yager wrote:

...
On Apr 1, 2023, at 09:57, MigMit wrote:

Well, human programmers can be shamed, yelled at, fired (which would actually hurt them), or, in extreme cases, prosecuted. They have every insentive to do their job right.

ChatGPT has RLHF. It has incentives to do its job right as well.

Serguey Zefirov

2:36 p.m.

What does reinforcement learning from human behavior reinforces exactly? The goal can be an engagement, which is the opposite of good behavior of a professional programmer. But my main concern with the code *produced* with ChatGPT is that the code is *produced*. The model is generative, not transformative. An anecdote: https://www.folklore.org/StoryView.py?story=Negative_2000_Lines_Of_Code.txt I've yet to see an example of ChatGPT reducing the size of code. If there are some, do their prompts still have directions to do so? The code is a liability - it has to be supported, maintained, kept, etc. I take a pride when I my changeset has more deletions than insertions, a rare occasion. 2023-04-01 17:09 GMT+03:00, Will Yager :

...

...
On Apr 1, 2023, at 09:57, MigMit wrote:

Well, human programmers can be shamed, yelled at, fired (which would actually hurt them), or, in extreme cases, prosecuted. They have every insentive to do their job right.

ChatGPT has RLHF. It has incentives to do its job right as well. _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Ivan Perez

3:37 p.m.

My two cents: - ChatGPT knows not just Haskell, but also specific libraries like Copilot ( https://github.com/Copilot-Language/copilot/) and the FRP library Yampa ( https://github.com/ivanperez-keera/yampa). Here's an example: https://www.facebook.com/photo?fbid=10161505593168357&set=a.427589328356 - At NASA we have systems (https://github.com/nasa/ogma) that help us transform structured natural language into Copilot. ChatGPT may help us go from unstructured NL to Copilot. We've tried it with properties like the ones we use in actual work and it produces something that is 95% of the way there. With the compiler for type checking, plus simulators and other tools for debugging, this could really accelerate the process. Obviously we'd need to check the result, etc. (which we already do regardless of who writes it). We are studying this direction atm. Ivan

Viktor Dukhovni

4:52 p.m.

On Sat, Apr 01, 2023 at 08:18:08AM +0200, Aloïs Cochard wrote:

...

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Part of the good laugh is on page 136: Let P be a point on the hyperbola x^2 + 3xy + 4x − 5y = −9y^2 − 133. Find the shortest possible distance from the origin to P. The authors of the paper say that "GPT-4 produces a sound argument", I beg to differ. Let u = 3y, then x^2 + xu + u^2 + 4x − (5/3)u = − 133. The degree-two part of which is positive-definite. The linear terms just shift the origin. So the equation is actually: r^2 + rs + s^2 = RHS For a straight-forward to compute choice of r = x - a, s = u - b. There are then two issues (just the first one is enough) with the prompt: * The equation can't represent a hyperbola, it would be an ellipse. * The ellipse doesn't exist, because the RHS constant is actually negative. THe authors are just as prone to autopilot nonsense reasoning as GPT-4. This rather reminds me of: https://math.stackexchange.com/questions/1594740/v-i-arnold-says-russian-stu... -- Viktor.

Aloïs Cochard

5:16 p.m.

Thanks for that Viktor, While I was holding my Giga Beer and looking at your delicious equations, I started solving them using Japanese Multiplication: https://www.youtube.com/watch?v=gngvWShRgX4 また会いましょう、そしてすべての魚に感謝します On Sat, 1 Apr 2023 at 18:53, Viktor Dukhovni wrote:

...

On Sat, Apr 01, 2023 at 08:18:08AM +0200, Aloïs Cochard wrote:

...
How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laugh

Part of the good laugh is on page 136:

Let P be a point on the hyperbola

x^2 + 3xy + 4x − 5y = −9y^2 − 133.

Find the shortest possible distance from the origin to P.

The authors of the paper say that "GPT-4 produces a sound argument", I beg to differ.

Let u = 3y, then

x^2 + xu + u^2 + 4x − (5/3)u = − 133.

The degree-two part of which is positive-definite. The linear terms just shift the origin. So the equation is actually:

r^2 + rs + s^2 = RHS

For a straight-forward to compute choice of r = x - a, s = u - b.

There are then two issues (just the first one is enough) with the prompt:

* The equation can't represent a hyperbola, it would be an ellipse. * The ellipse doesn't exist, because the RHS constant is actually negative.

THe authors are just as prone to autopilot nonsense reasoning as GPT-4.

This rather reminds me of:

https://math.stackexchange.com/questions/1594740/v-i-arnold-says-russian-stu...

-- Viktor. _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

Branimir Maksimovic

9:26 p.m.

Give it to solve this bellow one minute in Haskell: https://projecteuler.net/problem=512I have shown it solution, but anyway really doubt it is programmed to remember anything :P Greets, Branimir.

...

On 1. 4. 2023., at 18:52, Viktor Dukhovni wrote:

On Sat, Apr 01, 2023 at 08:18:08AM +0200, Aloïs Cochard wrote:

...
How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laughhi

Part of the good laugh is on page 136:

Let P be a point on the hyperbola

x^2 + 3xy + 4x − 5y = −9y^2 − 133.

Find the shortest possible distance from the origin to P.

The authors of the paper say that "GPT-4 produces a sound argument", I beg to differ.

Let u = 3y, then

x^2 + xu + u^2 + 4x − (5/3)u = − 133.

The degree-two part of which is positive-definite. The linear terms just shift the origin. So the equation is actually:

r^2 + rs + s^2 = RHS

For a straight-forward to compute choice of r = x - a, s = u - b.

There are then two issues (just the first one is enough) with the prompt:

* The equation can't represent a hyperbola, it would be an ellipse. * The ellipse doesn't exist, because the RHS constant is actually negative.

THe authors are just as prone to autopilot nonsense reasoning as GPT-4.

This rather reminds me of:

https://math.stackexchange.com/questions/1594740/v-i-arnold-says-russian-stu...

-- Viktor. _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Aloïs Cochard

10:38 p.m.

I'm sorry, I'm too busy adding constraints to the next version of Chat j'ai peter. On Sat, 1 Apr 2023 at 23:26, Branimir Maksimovic < branimir.maksimovic@gmail.com> wrote:

...

Give it to solve this bellow one minute in Haskell: #512 Sums of totients of powers https://projecteuler.net/problem=512 projecteuler.net https://projecteuler.net/problem=512 [image: apple-touch-icon.png] https://projecteuler.net/problem=512 https://projecteuler.net/problem=512 I have shown it solution, but anyway really doubt it is programmed to remember anything :P

Greets, Branimir.

On 1. 4. 2023., at 18:52, Viktor Dukhovni wrote:

On Sat, Apr 01, 2023 at 08:18:08AM +0200, Aloïs Cochard wrote:

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laughhi

Part of the good laugh is on page 136:

Let P be a point on the hyperbola

x^2 + 3xy + 4x − 5y = −9y^2 − 133.

Find the shortest possible distance from the origin to P.

The authors of the paper say that "GPT-4 produces a sound argument", I beg to differ.

Let u = 3y, then

x^2 + xu + u^2 + 4x − (5/3)u = − 133.

The degree-two part of which is positive-definite. The linear terms just shift the origin. So the equation is actually:

r^2 + rs + s^2 = RHS

For a straight-forward to compute choice of r = x - a, s = u - b.

There are then two issues (just the first one is enough) with the prompt:

* The equation can't represent a hyperbola, it would be an ellipse. * The ellipse doesn't exist, because the RHS constant is actually negative.

THe authors are just as prone to autopilot nonsense reasoning as GPT-4.

This rather reminds me of:

https://math.stackexchange.com/questions/1594740/v-i-arnold-says-russian-stu...

-- Viktor. _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

Mihai Maruseac

11:38 p.m.

At the risk of touting my own horn, I have been running experiments with chatGPT a while ago and both math (https://mihai.page/chatgpt/) and Haskell (https://mihai.page/chatgpt-2/) were still not there. I need to repeat these experiments though for the new version, these were run when GPT-3 was top of the line. On Sat, Apr 1, 2023 at 3:38 PM Aloïs Cochard wrote:

...

I'm sorry, I'm too busy adding constraints to the next version of Chat j'ai peter.

On Sat, 1 Apr 2023 at 23:26, Branimir Maksimovic < branimir.maksimovic@gmail.com> wrote:

...
Give it to solve this bellow one minute in Haskell: #512 Sums of totients of powers https://projecteuler.net/problem=512 projecteuler.net https://projecteuler.net/problem=512 [image: apple-touch-icon.png] https://projecteuler.net/problem=512 https://projecteuler.net/problem=512 I have shown it solution, but anyway really doubt it is programmed to remember anything :P

Greets, Branimir.

On 1. 4. 2023., at 18:52, Viktor Dukhovni wrote:

On Sat, Apr 01, 2023 at 08:18:08AM +0200, Aloïs Cochard wrote:

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laughhi

Part of the good laugh is on page 136:

Let P be a point on the hyperbola

x^2 + 3xy + 4x − 5y = −9y^2 − 133.

Find the shortest possible distance from the origin to P.

The authors of the paper say that "GPT-4 produces a sound argument", I beg to differ.

Let u = 3y, then

x^2 + xu + u^2 + 4x − (5/3)u = − 133.

The degree-two part of which is positive-definite. The linear terms just shift the origin. So the equation is actually:

r^2 + rs + s^2 = RHS

For a straight-forward to compute choice of r = x - a, s = u - b.

There are then two issues (just the first one is enough) with the prompt:

* The equation can't represent a hyperbola, it would be an ellipse. * The ellipse doesn't exist, because the RHS constant is actually negative.

THe authors are just as prone to autopilot nonsense reasoning as GPT-4.

This rather reminds me of:

https://math.stackexchange.com/questions/1594740/v-i-arnold-says-russian-stu...

-- Viktor. _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Mihai Maruseac (MM) "If you can't solve a problem, then there's an easier problem you can solve: find it." -- George Polya

Aloïs Cochard

2 Apr 2 Apr

6:12 a.m.

Absolutely no idea if there is an improvement, but this will should eventually help for such mathematical shenanigans: https://thenewstack.io/wolfram-chatgpt-plugin-blends-symbolic-ai-with-genera... Anyway, I hope everyone had a good April Fool's Day. https://en.wikipedia.org/wiki/April_Fools%27_Day On Sun, 2 Apr 2023 at 01:39, Mihai Maruseac wrote:

...

At the risk of touting my own horn, I have been running experiments with chatGPT a while ago and both math (https://mihai.page/chatgpt/) and Haskell (https://mihai.page/chatgpt-2/) were still not there.

I need to repeat these experiments though for the new version, these were run when GPT-3 was top of the line.

On Sat, Apr 1, 2023 at 3:38 PM Aloïs Cochard wrote:

...
I'm sorry, I'm too busy adding constraints to the next version of Chat j'ai peter.

On Sat, 1 Apr 2023 at 23:26, Branimir Maksimovic < branimir.maksimovic@gmail.com> wrote:

...
Give it to solve this bellow one minute in Haskell: #512 Sums of totients of powers https://projecteuler.net/problem=512 projecteuler.net https://projecteuler.net/problem=512 [image: apple-touch-icon.png] https://projecteuler.net/problem=512 https://projecteuler.net/problem=512 I have shown it solution, but anyway really doubt it is programmed to remember anything :P

Greets, Branimir.

On 1. 4. 2023., at 18:52, Viktor Dukhovni wrote:

On Sat, Apr 01, 2023 at 08:18:08AM +0200, Aloïs Cochard wrote:

How can this be useful when you have to anyway review everything is doing as he might to just randomly insert a bug or a security flaw??? I prefer to read poems by my human friends.

I highly recommend starting reading this paper at page 128 instead of wasting your time on that prompt: https://arxiv.org/pdf/2303.12712.pdf

Be ready for a good laughhi

Part of the good laugh is on page 136:

Let P be a point on the hyperbola

x^2 + 3xy + 4x − 5y = −9y^2 − 133.

Find the shortest possible distance from the origin to P.

The authors of the paper say that "GPT-4 produces a sound argument", I beg to differ.

Let u = 3y, then

x^2 + xu + u^2 + 4x − (5/3)u = − 133.

The degree-two part of which is positive-definite. The linear terms just shift the origin. So the equation is actually:

r^2 + rs + s^2 = RHS

For a straight-forward to compute choice of r = x - a, s = u - b.

There are then two issues (just the first one is enough) with the prompt:

* The equation can't represent a hyperbola, it would be an ellipse. * The ellipse doesn't exist, because the RHS constant is actually negative.

THe authors are just as prone to autopilot nonsense reasoning as GPT-4.

This rather reminds me of:

https://math.stackexchange.com/questions/1594740/v-i-arnold-says-russian-stu...

-- Viktor. _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- Mihai Maruseac (MM) "If you can't solve a problem, then there's an easier problem you can solve: find it." -- George Polya

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

David Feuer

31 Mar 31 Mar

9:29 p.m.

How does it do with explaining how purely functional catenable deques work? On Fri, Mar 31, 2023, 4:03 PM Gregory Guthrie wrote:

...

From a colleague:

GPT Pretty smart(?!).

But this is of course simple, and searchable.

Still…

--------------------------------------------------------------------------------------------------

GPT-4 when prompted:

*Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.*

gives the following response:

*-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element -- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned*

The above code runs:

*https://onecompiler.com/cpp/3z48yt2pp https://onecompiler.com/cpp/3z48yt2pp*

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

Aloïs Cochard

1 Apr 1 Apr

1:22 p.m.

Dear Gregory, Can you please fix your onecompiler vanity link? It points to the C++ runtime implementation. Thanks On Fri, 31 Mar 2023 at 22:04, Gregory Guthrie wrote:

...

From a colleague:

GPT Pretty smart(?!).

But this is of course simple, and searchable.

Still…

--------------------------------------------------------------------------------------------------

GPT-4 when prompted:

*Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.*

gives the following response:

*-- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element -- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned*

The above code runs:

*https://onecompiler.com/cpp/3z48yt2pp https://onecompiler.com/cpp/3z48yt2pp*

_______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

-- *Λ\oïs* http://twitter.com/aloiscochard http://github.com/aloiscochard

Gregory Guthrie

1:42 p.m.

Yes, will check on it. That is because the reason that a simple program was used was because the example was to ask GPT to solve it is a range of languages, one of which was Haskell and there were links to solutions in all of them.. “Write binary search in xxxx. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states.” { HTML, CSS and JavaScript, C++, Haskell, JavaScript, Python, CommonLisp, C, C#, PHP, Java, Go, Erlang, Alloy, TLA+, X-Machine, PVS Verifier, SPIN model checker, Guarded Command Language, Hoare triples, Gries & Owicki concurrent code} It was an interesting exercise from Dr. Nenad Koncar, a successful NL researcher and developer. I’ll ask him for a corrected link to the Haskell version. Dr. Gregory Guthrie From: Aloïs Cochard Sent: Saturday, April 1, 2023 8:23 AM To: Gregory Guthrie Cc: haskell-cafe@haskell.org; Nenad Končar ; Haskell Community Subject: Re: [Haskell-cafe] GPT & Haskell Dear Gregory, Can you please fix your onecompiler vanity link? It points to the C++ runtime implementation. Thanks On Fri, 31 Mar 2023 at 22:04, Gregory Guthrie mailto:guthrie@miu.edu> wrote: From a colleague: GPT Pretty smart(?!). But this is of course simple, and searchable. Still… -------------------------------------------------------------------------------------------------- GPT-4 when prompted: Write binary search in Haskel. Print results using putStrLn. Ensure bug-free code by specifying preconditions, postconditions, invariants, variants, internal states. gives the following response: -- Binary search in Haskell -- Precondition: the list is sorted in ascending order -- Postcondition: the function returns the index of the target element if found, or Nothing otherwise -- Invariant: the search space is always within the bounds of the list -- Variant: the search space decreases by half at each recursive call -- Internal state: the lower and upper bounds of the search space binarySearch :: Ord a => [a] -> a -> Maybe Int binarySearch xs x = go xs x 0 (length xs - 1) where go xs x lo hi | lo > hi = Nothing -- base case: search space is empty | x == midElem = Just mid -- base case: target element is found | x < midElem = go xs x lo (mid - 1) -- recursive case: search in the left half | otherwise = go xs x (mid + 1) hi -- recursive case: search in the right half where mid = (lo + hi) `div` 2 -- calculate the middle index midElem = xs !! mid -- get the middle element -- Print results using putStrLn main :: IO () main = do let xs = [1, 3, 5, 7, 9] -- a sorted list of integers let x = 5 -- a target element to search for let result = binarySearch xs x -- call the binary search function case result of -- pattern match on the result Nothing -> putStrLn "Element not found" -- print a message if Nothing is returned Just i -> putStrLn $ "Element found at index " ++ show i -- print a message with the index if Just i is returned The above code runs: https://onecompiler.com/cpp/3z48yt2pp _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post. -- Λ\oïs http://twitter.com/aloiscochard http://github.com/aloiscochard

826

Age (days ago)

828

Last active (days ago)

List overview

Download

38 comments

20 participants

participants (20)

Alexey Vagarenko
Aloïs Cochard
Brandon Allbery
Branimir Maksimovic
David Feuer
Dominik Schrempf
Gregory Guthrie
Ivan Perez
Jeff Clites
Jeffrey Brown
Levent Erkok
Mig Mit
MigMit
Mihai Maruseac
Serguey Zefirov
Tom Ellis
Viktor Dukhovni
Will Yager
William Yager
Yasuaki Kudo

GPT & Haskell

Gregory Guthrie

Gregory Guthrie

Yasuaki Kudo

Yasuaki Kudo

Tom Ellis

Gregory Guthrie

tags

participants (20)