CompleteSkeptic

4 ms·

1.
by CompleteSkeptic 10d ago
yes and can do many of those in parallel
2.
by CompleteSkeptic 10d ago
unfortunately all hand-written :( my chief-of-staff does unironically handwrite em dashes though
3.
by CompleteSkeptic 10d ago
the edit is right - jev would be cheaper, faster, and more self-consistent (in general) we actually use astra (and fable) in this way for our evals: evals.typesafe.ai someone on the team cooked hard on that and it shows example traces compa
4.
by CompleteSkeptic 11d ago
1. yes a general model 2. no training at all 3. but it is focused on "System 1" tasks (more human judgment, less math reasoning)
5.
by CompleteSkeptic 11d ago
strings (and all sequential data structures) are not allowed at all - this is how we make sure all outputs can be computed in parallel (thus no output token cost)
6.
by CompleteSkeptic 11d ago
They don't like adding stealth startups :(
7.
by CompleteSkeptic 11d ago
1. I am extremely on the same page 2. I do think that subconscious is not only much smarter than we give it credit for, but also much more robust than the "jagged frontier" of current LLMs (shilling my blog post on that jaggedness
8.
by CompleteSkeptic 11d ago
very accurate! the one nuance I'd get into is I'd call it "zero-shot" over "instruction-tuned" (the latter often implies a particular distribution), but very safe for sharing
9.
by CompleteSkeptic 11d ago
> 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?
10.
by CompleteSkeptic 11d ago
inputs are structured program state. there is an example at around second 30 of the doom demo (though ideally everyone gets off the waitlist and can try it out for themselves )
11.
by CompleteSkeptic 11d ago
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
12.
by CompleteSkeptic 11d ago
the hard part for coding is actually state engineering (e.g. getting your dependencies in context) - we haven't even tried it yet (because my philosophy is we should automate the easy tasks before the hard and we've been working o
13.
by CompleteSkeptic 11d ago
thanks a ton! constrained decoding (OpenAI-style structured outputs) make models dumber unfortunately - the short+dense version is that simply masking logits is insufficient because if ever a model was assigning probability to an invalid to
14.
by CompleteSkeptic 11d ago
architecture is close to the chest for now, but we have talked about writing a paper I don't want to shill my blog too much, but I will say data is probably far most interesting than architecture: https://www.completeskeptic
15.
by CompleteSkeptic 11d ago
love that you love the manifesto! letting the first batches off the waitlist now, but we do have some early users describing their experience ( https://x.com/danshipper/status/2099947471518474522 )
16.
by CompleteSkeptic 11d ago
that's right, but because these models are probabilistic, it's also possible to be confidently wrong (and all future models will be smarter still and still have that possibility)
17.
by CompleteSkeptic 11d ago
I'm biased but I wouldn't call it misleading - generating text is super awesome and flexible, (we describe that in the blog post - and I personally use string models all the time) but it's true you pay a high tax for autoregr
18.
by CompleteSkeptic 11d ago
exactly right!
19.
by CompleteSkeptic 11d ago
definitely not AI-generated - this is my real wardrobe we also thought the voice at the end was AI-ish, but apparently that's a real voice actor but slightly sped up
20.
by CompleteSkeptic 11d ago
we have played with this! the fascinating thing we've found so far is that adversarial examples for our model are quite different from that of LLMs so that they work even better together
21.
by CompleteSkeptic 11d ago
it is just a model, no harness yet ;) it is a structured data model, but technically not a language model (it doesn't generate language)
22.
by CompleteSkeptic 11d ago
it's our output tokens that are free (under the system one / jev column)
23.
by CompleteSkeptic 11d ago
you could, but it the model is not optimized for text this is complex, but generating text is highly complicated and requires mode dropping to make long cohesive text
24.
by CompleteSkeptic 11d ago
CEO here - that is right! I do agree that the comparison to LLM tokens is hard to understand (also because output tokens are not comparable). But yes, text or structured state (like a JSON with multiple pieces of text in) -> decisions ou
25.
(KV) Cache Rules Everything Around Me(completeskeptic.com)
3 points by CompleteSkeptic 17d ago|0 comments
26.
by CompleteSkeptic 3mo ago
constrained decoding tends to make models dumber - this is why it's rarely used
27.
Scaling Laws, Honestly(completeskeptic.com)
5 points by CompleteSkeptic 3mo ago|1 comments
28.
by CompleteSkeptic 4mo ago
I helped train some of the first "magic" models at OpenAI[1] and it was a wild ride. We were a pretty sane + skeptical team and we weren't totally convinced the models were as general as they seemed, but the query that convin
29.
by CompleteSkeptic 4mo ago
TypeSafe AI | Software Engineer | ONSITE (San Francisco) | Full-time TypeSafe AI ( https://typesafe.ai/ ) is a well-funded stealth mode company making a new class of AI models which enable a reliable way to embed composable i
30.
by CompleteSkeptic 5mo ago
Is this the first time OpenAI has published comparisons to other labs? Seems so to me - see GPT-5.4[1] and 5.2[2] announcements. Might be an tacit admission of being behind. [1] https://openai.com/index/introducing-gpt-

More ›