Hacker News Reader

Comment by jlawson | Hacker News Reader

Specs are ambiguous but not necessarily non-deterministic.

The same entity interpreting the spec in exactly the same way will resolve the ambiguities the same way each time.

Human and current AI interpretation of specs is non-deterministic process. But, if we wanted to build a deterministic AI we could.

View on HN

> But, if we wanted to build a deterministic AI we could.

Is this bold proposal backed by any theory?

Given that they all use pseudo-random (and not actually random) numbers, they are "deterministic" in the sense that given a fixed seed, they will produce a fixed result...

But perhaps that's not what was meant by deterministic. Something like an understandable process producing an answer rather than a pile of linear algebra?

I was thinking the exact same thing: if you don’t change the weights, use identical “temperature” etc, the same prompt will yield the same output. Under the hood it’s still deterministic code running on a deterministic machine

This is incorrect. Temperature would need to be zero to get same result.

This is not correct. If the algorithm is deterministic and the random seed is the same, temperature can be anything and get the same result.

Same as using a seed to get the same map generated in Dwarf Fortress.

Thanks for replying, this is what I thought as well.

You’re right - TIL

You can just change your definition of "AI". Back in the 60s the pinnacle of AI was things like automatic symbolic integration and these would certainly be completely deterministic. Nowadays people associate "AI" with stuff like LLMs and diffusion etc. that have randomness included in to make them seem "organic", but it doesn't have to be that way.

I actually think a large part of people's amazement with the current breed of AI is the random aspect. It's long been known that random numbers are cool (see Knuth volume 2, in particular where he says randomness make computer-generated graphics and music more "appealing"). Unfortunately being amazed by graphics and music (and now text) output is one thing, making logical decisions with real consequences is quite another.

In 2025, 99% of people are talking about LLMs or stable diffusion.

So then your question boils down to "how could X, which I've defined as Y, be Z?"

The cutting edge of AI research is LLMs. Those aren't deterministic.

You can build an "AI" with whatever you want, but context matters and we live in 2025, not 1985.

> if we wanted to build a deterministic AI we could

What's holding us back from building a deterministic generative AI?

I probably don't understand enough, but I assumed that the non-determinism was "inherit" in the current LLM technology

Reply on news.ycombinator.com