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

Promising results from DeepSeek R1 for code

simonwillison.net

Post reply on HN
Print view View original

441–450 of 765 posts

Re: Promising results from DeepSeek R1 for code

#441

Post by resource_waste »

My number 1 criticism of long term LLM claims is that we already hit the limit.

If you see the difference between a 7B model and a 70B model, its only slightly impressive. a 70B and a 400B model is almost unnoticeable. Does going from 400B to 2T do anything?

Every layer like using python to calculate a result, or using chain of thought, destroys the purity. It works great for Strawberries, but not great for developing an aircraft. Aircraft will still need to be developed in parts, even with a 100T model.

When you see things like "By 20xx", no, we already hit it. Improvements you see are mere application layers.

Re: Promising results from DeepSeek R1 for code

#442

Post by rahimnathwani »

anotherpaulg wrote: post #61

> 99% of the code in this PR [for llama.cpp] is written by DeekSeek-R1 It's definitely possible for AI to do a large fraction of your coding, and for it to contribute significantly to "improving itself". As an example, aider currently writes about 70% of the new code in each of its releases. I automatically track and share this stat as graph [0] with aider's release notes. Before Sonnet, most releases were less than…

When a log line contains {main_model, weak_model, editor_model} does the existence of main_model mean that mean the person was using Aider in Architect/Editor mode?

Do you usually use that mode and, if so, with which architect?

Thank you!

Re: Promising results from DeepSeek R1 for code

#443

Post by IncreasePosts »

catigula wrote: post #397

Dario Amodei says software engineering is fully automated by 2027. You might have the 0.01% engineer left over, but that's it, the job is finished. I think people need to start considering strongly what kind of career they can re-skill to. https://darioamodei.com/machines-of-loving-grace

What does that even mean?

What exactly is the .01% of engineering work that this super intelligent AI couldn't handle?

I'm not worried about this future as a SWE, because if it does happen, the entire world will change.

If AI is doing all software engineering work, that means it will be able to solve hard problems in robotics, for example in manufacturing and self driving cars.

Wouldn't it be able to create a social network more addictive than TikTok, for anyone who might watch? This AI wouldn't even need human cooperation, why couldn't it just generate videos that were addictive?

I assume an AI that can do ultra complex AI work would also be able to do almost all creative work better than a human too.

And of course it could do the work of paper shuffling white collar workers. It would be a better lawyer than the best lawyer, a better accountant than the best accountant.

So, who exactly is going to have a job in that future world?

Re: Promising results from DeepSeek R1 for code

#444

Post by qingcharles »

the_arun wrote: post #18
lawlessone wrote: post #9

Earlier quoted context omitted.

i spent a months salary on these eggs and can no longer afford to cook them :(

Hey, where are you getting the eggs? I am unable to find them in the market.

Sold my GPU, bought chickens.

Re: Promising results from DeepSeek R1 for code

#445

Post by catigula »

rileymichael wrote: post #425
catigula wrote: post #397

Dario Amodei says software engineering is fully automated by 2027. You might have the 0.01% engineer left over, but that's it, the job is finished. I think people need to start considering strongly what kind of career they can re-skill to. https://darioamodei.com/machines-of-loving-grace

gee, I wonder why the guy with an enormous vested interest in pushing this narrative would say that? in general, the people saying this sort of thing are not / have never been engineers and thus have no clue what the job _actually_ involves. seems to be the case here with this person.

Don't you think software engineers have a vested interest in their jobs being relevant, just with less information?

Re: Promising results from DeepSeek R1 for code

#446

Post by __MatrixMan__ »

throwu2728828 wrote: post #435
catigula wrote: post #397

Dario Amodei says software engineering is fully automated by 2027. You might have the 0.01% engineer left over, but that's it, the job is finished. I think people need to start considering strongly what kind of career they can re-skill to. https://darioamodei.com/machines-of-loving-grace

Re-skill to what ? Everything is going to be upturned and/or solved by the time I could even do a pivot. There's no point at all now, I can only hold onto Christ.

If you believe that everything will be solved by the time you can pivot, what will we need jobs for anyway? I mean, the bottleneck justifying most scarcity is that we don't have adequate software to ask the robots to do the thing, so if that's a solved problem, which things will remain that still need doing?

I don't personally think that's how it will go. AI will always need its hand held, if not due to a lack of capability then due to a lack of trust. But since you do, why the gloom?

Re: Promising results from DeepSeek R1 for code

#447

Post by yoyohello13 »

WXLCKNO wrote: post #280
submeta wrote: post #259

Earlier quoted context omitted.

The dev jobs won‘t go away, but they will change. Devs will be more and more like requirements engineers who need to understand the problem to then write prompts with the peoper context so that the llm can produce valuable and working code. And the next level will be to prompt llms to generate prompts for llms to produce code and solutions. But already I hire less and less developers for smaller tasks. The things tha…

I think what you're describing is going to be a very short transitional period. Once current AI gets good enough, the people micromanaging parts of it will do more to hinder the process than to help it. One person setting the objectives and the AI handling literally everything else including brainstorming issues etc, is going to be all that's needed.

> One person setting the objectives and the AI handling literally everything else including brainstorming issues etc, is going to be all that's needed.

A person just setting the prompt and letting the AI do all the work is not adding any additional value. Any other person can come in and perform the exact same task.

The only way to actually provide differentiation in this scenario is to either build your own models, or micromanage the outputs.

Re: Promising results from DeepSeek R1 for code

#448

Post by catigula »

throwu2728828 wrote: post #426
ldjkfkdsjnv wrote: post #400

Earlier quoted context omitted.

I fully believe this as well. And I have 15 years of SWE experience at top tech. Its over for this field

[flagged]

It's a shame that AI seems to be causing a lot of despair, even prior to its vision being complete.

I was forced to implement AI systems that toasted many of our employees.

Re: Promising results from DeepSeek R1 for code

#449

Post by rahimnathwani »

mike31fr wrote: post #362
simonw wrote: post #3

Given these initial results, I'm now experimenting with running DeepSeek-R1-Distill-Qwen-32B for some coding tasks on my laptop via Ollama - their version of that needs about 20GB of RAM on my M2. https://www.ollama.com/library/deepseek-r1:32b It's impressive! I'm finding myself running it against a few hundred lines of code mainly to read its chain of thought - it's good for things like refactoring where it will thi…

Noob question (I only learned how to use ollama a few days ago): what is the easiest way to run this DeepSeek-R1-Distill-Qwen-32B model that is not listed on ollama (or any other non-listed model) on my computer ?

This model is listed on ollama. The 20GB one is this one: https://ollama.com/library/deepseek-r1:32b-qwen-distill-q4_K...

Re: Promising results from DeepSeek R1 for code

#450

Post by kragen »

dboreham wrote: post #186
nkozyra wrote: post #114

Earlier quoted context omitted.

> it isn't entirely bluster "Development" is effectively translating abstractions of an intended operation to machine language. What I find kind of funny about the current state is we're using large language models to, like, spit out React or Python code. This use case is obviously an optimization to WASM, so a little closer to the metal, but at what point to programs (effectively suites of operations) just cut out t…

I've wondered about this too. The LLM could just write machine code. But now a human can't easily review it. But perhaps TDD makes that ok. But now the tests need to be written in a human readable language so they can be checked. Or do they? And if the LLM is always right why does the code need to be tested?

The LLM might be terrible at writing machine code directly. The kinds of mistakes I see GPT-4 making in Python, PostScript, or JS would be a much bigger problem in machine code. It "gets confused" and "makes mistakes" in ways very similar to humans. I haven't had a chance to try DeepSeek R1 yet.
Post reply on HN