You will notice that Gemini Carbon thinks differently than Claude and GPT. It often comes up with entirely different but more clever solutions than competitor models.
This is precisely because we never distilled other models. Distilling other models creates a poor imitation rather than unique intelligence.
@wafisher@SullyOmarr I don’t think this type of automation is appealing to most people.
Eating out is a reward, a treat for them. So they like the ritual of browsing (even to reorder the same thing).
There is value in even the illusion of control.
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had
with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works
https://t.co/zHij5zwKGT
Yeah it’s not that we can’t build planes like that anymore, but it’s not what people want to buy tickets for.
Ticket prices were just much higher then: a NY–LA round-trip around $150 in the early 1970s is roughly $1,000–$1,500 in today’s dollars, vs. often $200–$400 now.
Since 1978, the share of Americans flying in a given year went from about 25% to 40%+, and the share who’d ever flown rose from ~63% to 87–90%.
Denser seating and more carry-ons are the result of pricing pressure.
Asked Gemini 4 Argon for an interactive museum of Apple computers.
...and didn't provide any images or data.
It scraped archival Mac photos, wrote a tool to clean the images, pulled 50 years of specs, and built a 3D Cover Flow with UI that morphs to match the decade you’re in.
@Afinetheorem So true.
Reading these docs made me realize that, while these videos were “one shot”, considerable human craft and thought went into the initial prompt.
@HamelHusain This guy might work at a company where they don’t give him a VM to ssh into.
For entrepreneurs and startup people it seems obvious, but in big corps these simple things come with a ton of friction.
@brianbreslin@loobah_l Dental Hygienists typically earn more than a high school grad so they should be ok.
I’m with you on cosmetology grads though. If the government did more standard underwriting for repayment risk, that would be better.
@itsnotzoie@usedgov What will likely happen is colleges will come up with lower-cost tuition options for degrees with lower post-grad earning potential.
It’s not fair for colleges to charge more for tuition than a graduate can earn from the degree.
@danprimack@MattLira Yeah agreed.
It can really help affluent & busy people outsource more of these tasks by finding new avenues to hire paid help.
But for people with limited disposable income it s more likely to tell them all the tasks they need to personally complete.