Hacker News Reader

I think it’s less than reasonable to operate mainly on vibes, rumors, and hearsay.

You don’t need to think about climate change studies. Instead you can read the allegedly tainted studies we’re actually talking about and profess to all of us what is wrong with them. You can’t point to exactly where they’ve fudged them.

View on HN

In an alternate universe where all research is completely auditable and we all have infinite time, yes that's a valid approach. And you're welcome to spend your life going that route, but the rest of us are gonna trust our common sense on this.

They don't even need to fudge the data on the studies they have published. They just need to hold back other studies that contradict the idea. Pouring over the published data looking for flaws will never get us access to the unpublished data.

Even if this hypothesis that smaller specialized models are superior to larger models in a task as broad as software development, hyperscalers and the hyperfunded would still maintain a massive advantage in that they could just train 1000x more models than anyone else.

And let’s be real. If this were some conspiracy that OpenAI, Google, and Anthropic were tacitly or overtly conspiring on, I’m pretty sure Elon or Zucc would ruin it to get ahead

Reply on news.ycombinator.com