Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Code Llama, a state-of-the-art large language model for coding

ai.meta.com

Post reply on HN
Print view View original

521–525 of 525 posts

Re: Code Llama, a state-of-the-art large language model for coding

#521

Post by sam_goody »

chaboud wrote: post #505
tuukkah wrote: post #502

Earlier quoted context omitted.

This thread started with AI code whose correctness of is_prime no-one has questioned so far :-D -1 True 0 True 1 True 2 True 3 True 4 False

Well, I’m questioning it… not that it really matters. -1, 0, and 1 are no good. Two divisors for positive integer X: 1, X I recently demonstrated GPT-4 via by having it explain phases of matter metaphorically with animals. Though impressive, it points to loose abstraction being more in range than firm abstraction. Let’s get that model a python interpreter and see how far we can take this party. Until then, I’m going…

No need for the disclaimer on the dragons, especially as we can't prove they have not existed [even though we both know they haven't].

Re: Code Llama, a state-of-the-art large language model for coding

#522

Post by cosmojg »

EvgeniyZh wrote: post #180
nabakin wrote: post #138

Earlier quoted context omitted.

Looks like they aren't releasing a pretty interesting model too. In the paper they mention a "Unnatural Code Llama" which wipes the floor with every other model/finetune on every benchmark except for slightly losing to Code Llama Python on MBPP pass@100 and slightly losing to GPT-4 on HumanEval pass@1 which is insane. Meta says later on that they aren't releasing it and give no explanation. I wonder why given how inc…

Note that current GPT-4 pass@1 for HumanEval is closer to 90% than to 67% reported in GPT-4 technical report, as reported, e.g., in [1] [1] https://arxiv.org/abs/2305.01210

Unfortunately, we have no idea whether GPT-4 has been further trained or finetuned on contaminated data since then.

Re: Code Llama, a state-of-the-art large language model for coding

#523

Post by cosmojg »

zarzavat wrote: post #416
SparkyMcUnicorn wrote: post #207

Earlier quoted context omitted.

I'm not sure copilot is using codex anymore[0]. They've also been talking about a shift towards GPT-4 with "Copilot X" a few times now[1][2]. [0] https://github.blog/2023-07-28-smarter-more-efficient-coding... [1] https://github.com/features/preview/copilot-x [2] https://github.blog/2023-07-20-github-copilot-chat-beta-now-...

Copilot X is just their name for their project to bring AI to more areas of VSCode. I don’t believe they can use GPT-4 for completions because it’s a chat-optimized model. It seems that they are using something else, that blog post seems to imply it’s a custom-trained model.

I use GPT-4 for code completion all the time! There are many Neovim extensions[1][2][3] (and I'm sure there are many VSCode extensions) which call the GPT-4 API directly for code completion. I'm pretty sure the only reason that Microsoft might avoid using GPT-4 for Copilot is cost.

[1] https://github.com/cosmojg/nvim-magic

[2] https://github.com/dpayne/CodeGPT.nvim

[3] https://github.com/aduros/ai.vim

Re: Code Llama, a state-of-the-art large language model for coding

#525

Post by xcf_seetan »

astrange wrote: post #436
swader999 wrote: post #305

Earlier quoted context omitted.

The answer to AI stealing your job is to go ahead and start a company, solve a hard problem, sell the solution and leverage AI to do this.

The only thing that takes anyone's job is demand shortfalls. Productivity increases certainly don't do it. It's like saying getting a raise makes you poorer.

Actually, in my country, Portugal, if your salary is the Minimum wage, you are exempt from paying taxes, but if you get a raise as little as 15€ or so, you move one category up and start to pay taxes, and you will receive less money than what you used to get before the raise.
Post reply on HN