Thread

Beating GPT-5.6 Sol on retrieval with 100x cheaper open models

415 points · 114 comments · moonikakiss

cmiles8 · 1785978944 · focus · HN ↗
The big lab models are academically interesting but business wise they seem toast long term. There’s no way for these model companies to compete when the models are becoming a pure commodity and others offering options that are orders of magnitude cheaper.
It’s not that the big labs couldn’t theoretically just also put out 100x cheaper options but their business model requires them to generate huge revenues from higher priced tokens or they’ll implode.
1. kumama · 1785979924 · focus · HN ↗
  castform founder here. despite our bet on fine-tuned smaller open-source models, i'm still quite bullish on the big labs. i think scaled closed models will continue to dominate for more general purpose use-case like codegen, search, etc. but intelligence has lots of long-tail applications and i think for these longer-tail applications, finetuned custom models will rule
  1. sroussey · 1785980986 · focus · HN ↗
    Product search model:
    <a href="https://www.linkedin.com/posts/introducing-ontology-1-ugcPost-7488272626277294080-XxvU/" rel="nofollow">https://www.linkedin.com/posts/introducing-ontology-1-ugcPos...
    Edit: more direct links, sorry:
    <a href="https://onton.com/research/ontology-1" rel="nofollow">https://onton.com/research/ontology-1
    <a href="https://onton.com/research/ontology-1-benchmarks" rel="nofollow">https://onton.com/research/ontology-1-benchmarks
    kumama · 1785981193 · focus · HN ↗
    this is really cool. i'm sure some of the larger e-commerce companies are already using clickstream data to tune better query rewriters/ranking models
    podocarp · 1786009195 · focus · HN ↗
    Sorry but what is neurosymbolic model, it's not explained I think
    sroussey · 1786043427 · focus · HN ↗
    IMB YouTube video explainer:
    <a href="https://www.youtube.com/watch?v=ZfWDVO3rzeA" rel="nofollow">https://www.youtube.com/watch?v=ZfWDVO3rzeA
2. esafak · 1785983438 · focus · HN ↗
  Small models aren't going to take people's jobs. Agents using large models might. That's how their numbers make sense.
  1. cmiles8 · 1785986563 · focus · HN ↗
    Except there’s been very little evidence of this happening or about to happen, which is turning into a huge problem for the big labs.
3. typ · 1785985744 · focus · HN ↗
  Electricity becoming a commodity doesn't necessarily make nuclear reactors become commodities. And as we can see, the publication of open-weight models doesn't really diffuse the know-how of training SoAT models as we expected. Even the major cloud providers are still unable to build competitive ones from scratch on their own. That's very different than the traditional FOSS ecosystem, where everyone can copy, learn, and evolve on one's own (like, Bitcoin -> altcoins).
  On token pricing, I think it's very much bottlenecked by hardware (the aggregate of compute) rather than the number of competing models. Assuming that the ceiling of the token price is determined by the economic value a unit of compute can provide, then the less efficient ones would be priced out of the compute allocation. It's not consumers bidding up a limited number of different models, but more like tokens of different models bidding up the limited computing resource. Less-intelligent tokens (which are generated by weaker models) are crowded out by smarter tokens from the limited compute. My prediction is that we'll see a meaningful downward pressure on token prices only when the new batches of next-generation hardware get mass-deployed.
  1. podocarp · 1786008888 · focus · HN ↗
    I don't think the big companies not making Sota models being an indicator that open source failed. If all semiconductor patents expired tomorrow it would still be prohibitively difficult to manufacture your own CPU at home or at work.
    typ · 1786018565 · focus · HN ↗
    Me neither. But open weights are not open source in the first place. It wouldn't be surprising that the raw intelligence of an LLM model has little to do with the architecture (which is the part most published). The secret sauce, like how/what data gets curated and cleaned, has rarely been talked about. So, even if the labs upload their weights, competitors are very difficult, if not impossible, to replicate the technology independently.
  2. childintime · 1786026192 · focus · HN ↗
    > Electricity becoming a commodity doesn't necessarily make nuclear reactors become commodities.
    Bad comparison? Nuclear reactors become obsolete. The only reason to build a nuclear reactor is as a government vanity project, just to get the bribes you want, or weapons grade material. Dirty tech.
    LunaSea · 1786026849 · focus · HN ↗
    And you know the small benefit of having stable, large scale and low CO2 emission electricity production.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.