reissbaker

5 ms·

1.
by reissbaker 6d ago
I feel like you could use an open-source harness like Pi and get 100+% of what these closed APIs offer without getting locked to OpenAI. What do you think is missing from them?
2.
by reissbaker 2mo ago
No, OpenAI did not publish how they trained o1, and at the time there was significant misunderstanding and belief in the research community that they were using some kind of Monte-Carlo tree search. DeepSeek figured out GRPO on their own. S
3.
by reissbaker 2mo ago
It's even worse than that. From the article: > Open-weights models that don’t have dangerous capabilities are a public good Oh! And, uh, what's a "dangerous capability" according to Anthropic? Let's see, accord
4.
by reissbaker 2mo ago
I think it's pretty hard to hold that worldview: Anthropic couldn't ship a reasoning model until they copied DeepSeek R1's homework, and they've all copied DS-style super-sparse MoEs at this point too.
5.
by reissbaker 2mo ago
256+: GLM-5.2 (swap with Kimi K3 when it goes open-weight on Monday) <256: Actually, surprisingly, not a Chinese model but probably Laguna S2.1. The best Chinese model at this size is DeepSeek V4 Flash though <96: Qwen 3.6 27B <3
6.
by reissbaker 2mo ago
The cost to run Kimi is the cost of the GPUs (+ overhead of hiring humans for now to manage it). Kimi K3 does not change the demand curve for LLMs, it only changes the possible suppliers — and they're all competing for the same suppl
7.
by reissbaker 2mo ago
I would assume the opposite is true — with an open-weight Fable-class model, doesn't demand for GPUs go up ? Plenty of companies can now look at what Anthropic is offering — high per token costs for a very intelligent model — and do t
8.
by reissbaker 2mo ago
Previously they were a distant fourth. They're not going to single-shot catch up to OpenAI or Anthropic, but they moved up the ladder one rung. In the short term labs are not profitable, although supposedly Anthropic is close. But Amaz
9.
by reissbaker 3mo ago
Agreed this sucks. We publish ours here and try to be as transparent as possible: https://synthetic.new/rate-limits
10.
by reissbaker 3mo ago
Interesting: I don't see anything in our error logs but we could be missing something (and personally the chat works for me + my unsubscribed test account). If you email us at hi@synthetic.new though we should be able to fix anything y
11.
by reissbaker 3mo ago
I'm biased because I run an inference company, https://synthetic.new . That being said I think we're pretty good at serving at GLM-5.2 — and other models, like Kimi K2.7! — and our privacy policy is quite good: zero dat
12.
by reissbaker 3mo ago
Self-promo but you should try our service synthetic.new. We generally have up-to-date open-source LLMs on the sub, and we have GLM-5.2 :) Perf+stability should be wayyy better than zai.
13.
by reissbaker 3mo ago
I wonder if this explains why I hear about this more from Europeans than from the SF tech scene. California is at-will employment, so you can fire an employee as easily as a contractor. Ironically this makes companies more willing to hire a
14.
by reissbaker 3mo ago
The Wikipedia articles say the majority of scholars believe it's based on Aramaic, while a minority of people (primarily non-linguistic-specialists in India) disagree. I think you're the one drawing from bias.
15.
by reissbaker 3mo ago
No, modern Hebrew and ancient Hebrew mapped similarly well to the written script — the primary difference between the two is just consonant drift. Both used the same structure of triconsonant roots with affixed patterns, and modern Hebrew m
16.
by reissbaker 3mo ago
As per the Wikipedia links, it's generally considered by scholars to be the origin of all alphabets and an early alphabetic script. Abjad is a term invented in 1990 to distinguish early alphabetic scripts without vowels from later scri
17.
by reissbaker 3mo ago
Hebrew is not based on Yiddish, lol; only Ashkenazi Hebrew pronunciation was influenced by Yiddish. Modern Israeli Hebrew uses primarily Sephardi pronunciation, and Ashkenazi is mocked (i.e. Shabbat is Sephardi, Shabbos is Ashkenazi; modern
18.
by reissbaker 3mo ago
Egyptian heiroglyphs were not an alphabet, even if they had alphabetic elements (in addition to pictographic ones). Scholars generally agree that proto-Sinaitic was the first alphabet, and all subsequent alphabets used today are either dire
19.
by reissbaker 3mo ago
No: most scholars believe alphabets were only invented once, much like the wheel. All Western alphabets are direct descendants, and the non-Western alphabets were directly inspired by it. [1] Phonetic alphabets were introduced to most of As
20.
by reissbaker 3mo ago
It wasn't directly cribbed (unlike Western alphabets), but given that Hangul was invented in the 1400s after exposure to Western alphabets, most scholars still consider alphabets to have only been invented once [1] and then copied, m
21.
by reissbaker 3mo ago
Fun fact: all (non-Cherokee?) alphabets in use today stem from an ancient Canaanite alphabet called the proto-Sinaitic script [1]. This is why Hebrew's alphabet near-perfectly phonetically represents the spoken language: Hebrew is just
22.
by reissbaker 3mo ago
America has somehow managed to hang on to the right to encryption, despite plenty of well-heeled opponents, so it's possible to hang on to the right to open-source models. But it'll take a lot of vocal support, since there'
23.
by reissbaker 3mo ago
You don't need to be able to self-host it. It's fine to pay someone else for it. If it's open-source, competition will ensure inference providers support it well enough, and if an open-source provider is dumb enough to nerf t
24.
by reissbaker 3mo ago
Many, many, many public policy positions; for a clear-cut example, they eventually supported SB 1047 [1] which would have banned open-sourcing any model trained with over 10^26 FLOPS (i.e. what Anthropic reportedly used to train Mythos). Th
25.
by reissbaker 3mo ago
The same way Anthropic is making it difficult to compete with them. They intentionally train the model (via PEFT, as called out in the model card) to be dumber when attempting to do things Anthropic doesn't want — in this case, competi
26.
by reissbaker 3mo ago
"You can't take code produced by our service to make competing services, but we can take code you produced to compete with your service (i.e. software engineering)" is pulling up the ladder IMO. If they can from-scratch train
27.
by reissbaker 3mo ago
They trained their models on everyone's data on the internet, and certainly violated many website terms of service.
28.
by reissbaker 3mo ago
It's becoming extremely important to support open-source AI, especially legally. Anthropic is willing to go totalitarian this quickly; imagine how much worse they'd be willing to do with government-granted monopolies that ban op
29.
by reissbaker 4mo ago
We've come full circle: now we have AI-written blog posts complaining about AI-written code.
30.
by reissbaker 4mo ago
Qwen 3.6 27B is fine but it's not in the same ballpark as GLM-5.1 or Kimi K2.6. If you truly want to scale up, you should get the 8xH200 with NVLink.

More ›