HackerTrans
トップ新着トレンドコメント過去質問紹介求人

par1970

no profile record

投稿

Ask HN: Are the newest LLMs better than you at programming?

4 ポイント·投稿者 par1970·4 か月前·19 コメント

コメント

par1970
·先月·議論
> AI (in this form) will never be able to solve things we truly cannot solve yet.

Argument?
par1970
·2 か月前·議論
> LLMs are nothing close to AGI and not going to lead to it, they can’t distinguish right from wrong, they can’t count, they can’t reason, they generate plausible text from a vast databank of connected text.

Argument?

Are LLMs close to being able to significantly help AGI researchers?
par1970
·2 か月前·議論
Which doomer argument have you found what problem with?
par1970
·3 か月前·議論
Do you deny the reported bug finding capabilities, or do you deny that they are dangerous?
par1970
·3 か月前·議論
I think this is roughly solved. Tell the agent to do all of its calculations in python.
par1970
·3 か月前·議論
Do you have a defense of why human-hammer-nail is a good analogy for human-chatgpt5.4-pwndsamsung?
par1970
·3 か月前·議論
But the service also tells criminals and adversaries about the bomb locations.
par1970
·3 か月前·議論
Yeah, maybe you are right. But is doing math and reasoning about Turing machines a priori? If so, then it seems plausible to me that reasoning about a codebase (without running it) is also ‘a priori’.
par1970
·3 か月前·議論
> What do you mean "a priori understanding codebases"?

I took him to be distinguishing between (1) just reading the code/docs and reasoning about it, and (2) that + crafting and running tests.
par1970
·3 か月前·議論
Why?
par1970
·4 か月前·議論
Did you tell it that it should test, or did you have it generate actual tests that you could run if you wanted to?
par1970
·4 か月前·議論
How much domain experience do you have? Is it helping you solve problems for paying customers?
par1970
·4 か月前·議論
If your project requires the solution of a tricky algorithmic issue, then is the AI system able to solve that part, or do you have to give it the solution?
par1970
·4 か月前·議論
What models + versions are you using?

Is it bad at designing systems that don't have a bunch of integrations?
par1970
·4 か月前·議論
> I don't use ChatGPT, but i've been using an agent with Claude Sonnet 4.

Are you using Sonnet 4.6?

> So this AI Agent... It is much faster at doing code when given specific instructions. But it keeps loosing context on architecture, and i cant really let it build complex things with interdependencies that build on each other.

I've only built small things (< 1000 lines) with the systems, so I might be missing this problem.

Is it better than you at building small self-contained things?

> And i get a bad feel when i then wonder how is this app doing what it does? because my agent cant explain it, and i would be stupid to believe what it hallucinated because it sounds really solid until you scratch the construction.

Do you ask it to generate test suites for the things that it builds?

> it would be also faster to build a catastrophic spaghetti code nightmare if not used with great care.

noted
par1970
·4 か月前·議論
Which models + versions are you using? Can you give a specific problem that you found them to be bad at?
par1970
·8 か月前·議論
Maybe we are talking past one another.

> Right, but that's explicitly not the body of government meant to represent people.

I haven't claimed that the Senate was intended to represent the people. I also haven't claimed that OP claimed that the Senate was intended to represent the people.

> So is he saying the Senate is fundamentally a ridiculous way of representing 100 states, or is he saying the House is fundamentally a ridiculous way of representing 350 million people?

He didn't say either of those things. He said this "The Senate is fundamentally a ridiculous way of representing 350 million people."
par1970
·8 か月前·議論
I think OP is arguing that because they literally said "The Senate is fundamentally a ridiculous way of representing 350 million people and we’re going to continue to get absurd unrepresentative outcomes for as long as it remains a relevant body."

What do you think they are arguing?
par1970
·8 か月前·議論
Are you arguing this?

(Premise 1) If a country has 350 million people, then the Senate will produce unrepresentative outcomes.

(Premise 2) America has 350 million people.

(Conclusion 1) So, the Senate will produce unrepresentative outcomes in America.

(Conclusion 2) So, the Senate is bad for America.
par1970
·8 か月前·議論
> We're nowhere close to AGI and don't have a clue how to get there.

Do you have an argument?