I think part of the battle is actually just getting people to identify which LLM made it to understand if someones contribution is good or not. A javascript project with contributions from Opus 4.6 will probably be pretty good, but if someone is using Mistral small via the chat app, it's probably just a waste of time.

But if I ask Claude or Gemini for a nice version of the recipe, it works perfectly. I think there's a lot of own goals out there.

Imagine the quality of life upgrade of getting compaction down to a few second blip, or the "Explore" going 20 times faster! As these models get better, it will be super exciting!

But to be clear, 1000 tokens/second is WAY better. Anthropic's Haiku serves at ~50 tokens per second.

(I can also see a world where it just doesn't make sense to share most of the layers/infra and you diverge, but curious how you all see the approach.)

Also, I put together a little research paper recently--I think there's probably an underexplored option of "Use frontier AR model for a little bit of planning then switch to diffusion for generating the rest." You can get really good improvements with diffusion models! https://estsauver.com/think-first-diffuse-fast.pdf

username at gmail if you want to chat

Claude Code v2.1.37

EU region, Claude Max 20x plan

Mac -- Tahoe 26.2

In the early years, it was extremely, extremely open and comprehensive. I've definitely looked through it when I wasn't sure how to handle something at work.

Put up a big "This property is not for sale" sign on the land.

https://estsauver.com/blog/scaling-visualizations

That being said, I don't think I would want to rent a car that didn't have a place to charge it or a very easy-to-use fast charger nearby.

Comments URL: https://news.ycombinator.com/item?id=46619954

Points: 2

# Comments: 0

> So you are saying that these special hybrid seeds that are the first generation of combining two strains are the only ones that can perform well?

For a lot of crop systems, yes! There are obviously crop systems where you can do clones and some exceptions are always present in biology, but basically yes. Yes for all the big staple crops except Canola.

> And that using any other seeds, even the second generation of that same strain, is so bad and so easy to confuse that it should be outright illegal?

I probably wouldn’t make it illegal, I think farmers should be allowed to do whatever they want to! (My completely out of the loop guess is the government is trying to help small holder farmers who are reporting that they’re being scammed by these groups and that they lack the resources to do genetic testing to prosecute them for the fraud.)

> That is very hard to believe. Such laws are in place to protect the IP of these special seed producers, to make their business model viable. That does have merit to a degree, you do want such companies to exist, but they should also have to contend with competition from other, perhaps less effective but cheaper, sources of seed.

It’s not really an IP protection thing, it’s an extremely difficult many year process to recover genetics on most hybrid crop systems. I don’t think most seed companies care about folks using saved seed, they know almost all farmers will buy good seed if they can.

> This doesn’t have much to do with protecting the farmers from being cheated into planting bad seed. And I am skeptical of the fact that even second generation seeds are that bad, or that these hybrids are really such a life-changing upgrade.

I think well answered by a parent comment, but the book The Wizard and The Prophet is pretty good reading on Borlaug and the green revolution. If you look at global food capacity vs population, it’s probably the single most important life upgrade for everyone of modernity.

(Small Edit: I should note that I’m not an agronomist, I’m just a guy who codes okay sometimes and that I’ve gotten to spend a lot of time talking to agronomists and smallholder farmers trying to make agriculture for small farmers work better.)