moab

Rochus, a huge thanks for the comment and discovering your fascinating site. I was blown away by how musical the recent Suno covers of your funk songs are. I would not have been able to guess that they were AI generated. Please post more videos in the future---I am convinced there is a large audience that would be interested in learning more about your setup and seeing you improvise.

I love it when people push back like this without suggesting the real reason. Can't think one can you? That too from a new sock puppet.

Bro, are we reading the same book? The book is totally uncritical of the subject and paints him like the second coming of christ. It feels like GDM wanted a canonization of Hassabis, and the writer simply obliged. Also, how does everything that GDM did keep coming back to some vague ideas in the guy's thesis? He is a great leader, no doubt, but him winning the Nobel Prize was just a huge joke.

Out of all the heads of AI orgs out there, Dennis is the best, but the book did him a disservice by painting an unrealistically sunny picture of him as some kind of visionary figure.

I'm not sure I agree. They're just essay prompts. One could write a bad essay that takes itself too seriously given the prompts, but one could also write a powerful essay starting from any of these prompts. I don't really see where you get the conclusion that students matriculating from these places have recently begun to be smoke-blowers that while possessing detailed knowledge of various arcana fail to produce anything useful.

Is anyone surprised? I'm reminded of how I felt during the NFT craze. LLMs are extremely powerful when used with deliberate care. Gas Town is the exact opposite of what is needed to actually do useful things in prod. I guess good on Steve for doing what he does so well, and getting so much hype around a vibe coded mess.

So to verify their claims and see how strong these models are, the answer is "believe us"?

Note: I'm expressing some skepticism here largely due to how recent rollouts from Meta flopped. Sincerely hoping that they do better this time around!

"Muse Spark is available now, and Contemplating mode will be rolling out gradually in meta.ai."

How does one get their hands on these models? They are not open-source, right? I go to meta.ai, but it's just a chat interface---no equivalent to codex or claud code? Can you use this through OpenCode? Is meta charging for model access, or is the gathering of chat data a sufficiently large tithe?

No. Read the book "Careless People". Meta leadership tried to downplay it by saying the stories are exaggerated. It seems doubtful to me.

I hope "OpenAI" gets the proverbial sword in the nuts once we get a change of government in this country. Probably unrealistic to hope for. Can a company be more hypocritical after openly bribing the pedophile in charge of this country?

How does Brockman sleep at night? These guys used to seem like standup ethical guys. It seems no amount of intellectualism is enough to ward off the poison of wealth.

Are you a professor at a R1 school? All the faculty I know at R1s (see CMU, MIT, etc) are doing quite a lot of teaching in addition to their research.

Let me use the discussion of coffee to plug what I believe is the most beautiful and long-lasting espresso machine currently out there (you'll hand it down to your grandkids): https://coffeegeek.com/reviews/firstlooks/cafelat-robot-espr...

I've pulled a few thousand cups of long espresso from this guy since we bought it two years ago. Much, much nicer and lower maintenance than a boiler machine. If one wants, you can go deep down the rabbit hole of heat control, etc. but even as a "just boil water and make espresso" machine it works great, with no fuss.

Unfortunately, after buying this thing I can't justify buying other coffee objects that are beautiful but would probably make worse coffee than the robot, e.g., the Moccamaster and other drip machines.

This is a bad take. The article makes it clear that most of them will lose money on the venture, and the reason the prices are high are due to status-mining chinese elites and traditional-medicine paranoiacs in vietnam. It's a pretty dismal situation.

The point stands that it's less about the language than doing said hard work in any reasonable programming language.

This is a very real worry for the AI rollout for the general population. But are folks here using AI to blow smoke up their asses as a sibling comment stated? I'd like to believe we're using it to ask questions, prototype, and then measure... not just blow smoke up there...

Napkin math is how you avoid spending several weeks of your life going down ultimately futile rabbit holes. Yes, it's approximations, often very coarse ones, but done right they do work.

Your question about what degree of parallelization is unfortunately too vague to really answer. SSDs offer some internal parallelism. Need more parallelism / IOPS? You can stick a lot more SSDs on your machine. Need many machines worth of SSDs? Disaggregate them, but now you need to think about your network bandwidth, NICs, cross-machine latency, and fault-tolerance.

The best engineers I've seen are usually excellent at napkin math.

You can do that by mowing, fyi.

You're entitled to your own opinion, but imo the point of posting anything on HN is to subject yourself to feedback. That's what I gave. Feedback.

How about not killing the weeds? One doesn't need to live a perfectly manicured pesticide-ridden hellscape.

Do you know of important real-world use-cases where cache-oblivious data structures are used? They are frequently mentioned on HN when relevant discussions like this one pop up, but I would love to hear about places where they are actually used in production.

You clearly did not look at the proposed cuts for the coming years---it guts the NSF.

Giuseppe, any interesting comments from inside the tent? From the outside, spending 100M+ on a single researcher seems like a wild idea. It's also not clear what spending that much money on people at the top of the leadership stack really brings that Meta does not already have in its existing leadership. Surely what is needed are strong hands-on researchers? Are they paying 100M a head for such people? If so, the times are truly wild.

It looks like just a wrapper around the data from plasticlist for now. One can fund other products, but I searched and could not find any others that were funded as a result of this project. Some transparency about the cost seems critical for successfully running such a crowd-funding project.

Yes, but that can be done once at ingestion time. Then retrieval is done over the pre computed FDEs using MIPS.

Watt-hours is the only unit that makes sense here as we're describing the energy cost of some calculation. Watt, a rate unit, is not appropriate.

Reading a book is like mindfulness meditation? Man, we really are living in a degraded age.

Why are we submitting AI linkspam about the oligarch class to HN?

Yes, let’s see how well the research kowtows to the party line. A careful, scrutinizing second review should do no harm, right?

> "OpenBrain (the leading US AI project) builds AI agents that are good enough to dramatically accelerate their research. The humans, who up until very recently had been the best AI researchers on the planet, sit back and watch the AIs do their jobs, making better and better AI systems."

I'm not sure what gives the authors the confidence to predict such statements. Wishful thinking? Worst-case paranoia? I agree that such an outcome is possible, but on 2--3 year timelines? This would imply that the approach everyone is taking right now is the right approach and that there are no hidden conceptual roadblocks to achieving AGI/superintelligence from DFS-ing down this path.

All of the predictions seem to ignore the possibility of such barriers, or at most acknowledge the possibility but wave it away by appealing to the army of AI researchers and industry funding being allocated to this problem. IMO it is the onus of the proposers of such timelines to argue why there are no such barriers and that we will see predictable scaling in the 2--3 year horizon.

It's a real pity they got discontinued. The sleepbuds work fantastic and are the only product that works for me.