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

Introducing System One Models and Jev

typesafe.ai

Post reply on HN
Print view View original

321–330 of 513 posts

Re: Introducing System One Models and Jev

#321

Post by qingcharles »

wg0 wrote: post #319
cfowles wrote: post #268

Wasn't really till seeing this home assistant demo they have ( https://www.loom.com/share/18c4dbcf8db546dfb2d7f2ef018e78e4 ) that the value really clicked for me. Seems really cool.

That's good. Side note - just like most people don't need an intelligent personal assistant to manage and respond their emails and book their flights, most people also don't need smart homes. Century old toggle switches are more than enough in a 3 room apartment or 5 room house unless you have a mention.

My primary beef with smart home (having tried it) is that every person that visits your home ends up confused about some element of it. A light switch that goes up and down is universally understood.

Re: Introducing System One Models and Jev

#322

Post by copperx »

wg0 wrote: post #317

Can I put it as Air Traffic Controller? With similar error rates as humans? That would be the litmus test. "Does not hallucinate" is not the same as "is never wrong". So the ATC test could be the benchmark.

Not hallucinating is easy when you don't produce strings.

Re: Introducing System One Models and Jev

#323

Post by copperx »

iforgotmypasswo wrote: post #290
skerit wrote: post #41

So in theory you could feed it incomplete text, and then ask it for the probabilities of what the next character could be?

Or a partially completed song, asking for the next note. I’m not sure if you’re joking, but using it for space constrained next token generation within a grammar sounds like a really neat use case.

Feed the generated note back into the input for the next query and you have ... autoregression?

Re: Introducing System One Models and Jev

#324

Post by copperx »

jawns wrote: post #38

I could see this being fantastic for classification tasks. Last year I shifted from using LLMs for bulk data classification tasks (1M transcripts) to generating embeddings and categorizing based on cosine similarity. It saved a ton of costs and time, but wasn't as accurate as LLMs. This seems like it can give me Terra-level classification ability with the cost/speed I need.

This ought to work better than SpamAssassin, I'm sure

Re: Introducing System One Models and Jev

#325

Post by bobtheborg »

copperx wrote: post #199
CompleteSkeptic wrote: post #124

Earlier quoted context omitted.

we hope so! the bigger hope is to not just eat LLM market share, but to allow for people to use AI much more in the inner loop of software

I'm sure you've thought of self-driving. How does the model work in that space?

Great question! Yes, this works much like the doom player. Sensor data (LIDAR, velocity, etc.) becomes the state. You use the score primitive to operate the controls ("What level of braking should be applied" 0: None, 1: just slightly slowing down, 2: there's a suspicious cat on the side of the road you don't trust, ...

Full disclosure, I am not they :=)

Re: Introducing System One Models and Jev

#326

Post by copperx »

wesammikhail wrote: post #288
bigglebear wrote: post #258

Earlier quoted context omitted.

I would guess a tiny stripped down text diffusion model. It only has 32k context, and for choice mode it can only select from 10 choices.

> and for choice mode it can only select from 10 choices. Rip there goes my excitement. I have a task that something like this would be great for but the list of options is a zero or two larger than that xd

They said that it works with up to 255 options.

Re: Introducing System One Models and Jev

#327

Post by wg0 »

lubujackson wrote: post #93

After much fumbling around with prompts and evals, this is exactly how I am using LLMs in production, to narrowly make choices and return structured data. Any deterministic work gets pulled out of the prompt and my goal is to narrow the model output to be as clearly defined and as minimal as possible. Jev's focus on structured I/O and confidence scores are game changing. If this does at all what it claims, I think th…

Curious what your use case is if not confidential.

Re: Introducing System One Models and Jev

#328

Post by monkpit »

bigglebear wrote: post #247
CompleteSkeptic wrote: post #141

Earlier quoted context omitted.

> Type safety is not factual correctness. I very much agree with this and want to hone in on where do actually disagree. Would you say a linear classifier hallucinates?

User input: "Hey, have your human support agent call me, tomorrow at 5pm." Model input: "Does the user want to speak to a human support agent?" Output: Yes. I imagine that your model would produce this, and I think it's fair to say this is a hallucination. A human would caveat it with: "Yes, but not right now.", your model is incapable of that. Yes is technically correct, but within the context of being in a live cha…

To be fair - you’re crafting a deliberately bad model input for a contrived example.

Re: Introducing System One Models and Jev

#329

Post by wg0 »

copperx wrote: post #199
CompleteSkeptic wrote: post #124

Earlier quoted context omitted.

we hope so! the bigger hope is to not just eat LLM market share, but to allow for people to use AI much more in the inner loop of software

I'm sure you've thought of self-driving. How does the model work in that space?

But the real problem in self driving isn't the decision making but object description. That is, computer vision if with cameras.

Decision making isn't that of a bottleneck I suppose.

Re: Introducing System One Models and Jev

#330

Post by monkpit »

zenlikethat wrote: post #215
seizethecheese wrote: post #123

Earlier quoted context omitted.

Just to be sure that I understand, you're saying that your model "can't hallucinate" because it only outputs a single thing, right? In this way, an LLM can't hallucinate either if I prompt it to do a classification task with a discrete set of possible outputs, right? (Assuming I reject non-conforming output. Actually, maybe what you're saying is that your system can't output non-conforming output?)

Yeah that's precisely correct. For e.g. classification tasks, even in 2026 people are doing things like hallucinating deliberately, and then matching the hallucinations to embeddings - https://softwaredoug.com/blog/2026/08/10/hypothetical-classi ... With TypeSafe it just picks the class (actually probabilities across classes), reliably every single time.

404?
Post reply on HN