msvana

Joined 3 years ago

Karma

Submissions

Member since

Aug 21

AI and software engineer. Interested in far too many things, including but not limited to AI, software development, economics, philosophy, or psychology.

https://svana.name

View on Hacker News

Recent submissions

Comment · 7 months ago

I think there is one problem with defining acceptance criteria first: sometimes you don't know ahead of time what those criteria are. You need to poke around first to figure out what's possible and what matters. And sometimes the criteria are subjective, abstract, and cannot be formally specified.

Of course, this problem is more general than just improving the output of LLM coding tools

Comment · 7 months ago

I work as an ML engineer/researcher. When I implement a change in an experiment it usually takes at least an hour to get the results. I can use this time to implement a different experiment. Doesn't matter if I do it by hand or if I let an agent do it for me, I have enough time. Code isn't the bottleneck.

I also heard an opinion that since writing code is cheap, people implement things that have no economic value without really thinking it through.

I built an RGB controller with Arduino svana.name

9 msvana 8 months ago 1 comments

Do LLMs hallucinate more in Czech than in English? svana.name

2 msvana 8 months ago 0 comments

Hallucinations in LLMs: What are they and what causes them svana.name

1 msvana 9 months ago 1 comments

Managing GPU Rentals with Rsync: Workflow for Volatile Cloud Resources svana.name

2 msvana 1 year ago 0 comments

Implementing a local AI coding agent is hard svana.name

2 msvana 1 year ago 1 comments

Show HN: Read-only AI coding assistant github.com

1 msvana 1 year ago 0 comments

How I solved PyTorch's cross-platform nightmare svana.name

73 msvana 1 year ago 29 comments

Lessons from AI Safety for Businesses svana.name

1 msvana 1 year ago 0 comments

Ask HN: How do you imagine an AI-driven utopia?

1 msvana 1 year ago 2 comments

Comment · 1 year ago

This reminds me of the idea that LLMs are simulators. Given the current state (the prompt + the previously generated text), they generate the next state (the next token) using rules derived from training data.

As simulators, LLMs can simulate many things, including agents that exhibit human-like properties. But LLMs themselves are not agents.

More on this idea here: https://www.alignmentforum.org/posts/vJFdjigzmcXMhNTsx/agi-s...

This perspective makes a lot of sense to me. Still, I wouldn't avoid anthropomorphization altogether. First, in some cases, it might be a useful mental tool to understand some aspect of LLMs. Second, there is a lot of uncertainty about how LLMs work, so I would stay epistemically humble. The second argument applies in the opposite direction as well: for example, it's equally bad to say that LLMs are 100% conscious.

On the other hand, if someone argues against anthropomorphizing LLMs, I would avoid phrasing it as: "It's just matrix multiplication." The article demonstrates why this is a bad idea pretty well.