Just this week Secretary Wright visited LLNL and while he made Ai as the next Manhattan project for the national labs. I got this feeling the next project for him was to cut back many of the programs.
Subscribe to:
Post Comments (Atom)
Posts you viewed tbe most last 30 days
-
Physics Today has an article about what went wrong with the LANL/LLNL contract. http://scitation.aip.org/content/aip/magazine/physicstod...
-
Did George W. Bush lie about weapons of mass destruction in Iraq ? Not looking for controversy here, just some facts from former LLNL/LANL e...
-
The budget has a 20% decrease to DOE office of science, 20% cut to NIH. NASA also gets a cut. This will have a huge negative effect on the ...
9 comments:
AI is the only thing labs should be doing. Just like the Manhattan project AI should first be developed and built at the labs. AI is brand new and the future is now.
Maybe AI is the next Manhattan Project, but why do it at the labs? That makes more sense at OpenAI, xAI, Palantir or Google. I don't doubt the importance of AI, but government needs to have more than a one track mind. We are falling behind the Chinese in so many areas of national defense. AI is just one component of that. Sometimes AI is needed as a critical piece of the puzzle, sometimes it plays a supporting role. We need to solve problems, not pursue trends for their own sake. Hammers looking for nails.
“We need to do AI at the labs….”
Yeah….I’m pretty sure we can get that project mission statement crafted in about a decade!
Los Alamos missed the boat on the modern nuclear weapon in the 1950’s which was developed by the likes of Teller, Johnny Foster, Seymour Sacks, and Harold Brown. I doubt LANL will succeed on getting on the AI bandwagon. But hey, they’ll always have “work life balance”.
While some people were daydreaming about what the Secretary was saying about Ai he went into great lengths about how he wants to streamline the labs, reduce bureaucracy, reduce redundancy, reduce overhead. Just quite a bit of hint of upcoming reduction in that brief. One interesting thing that was mentioned briefly was LANL and LLNL were out of space but INL had quite a bit of open space that isn’t being used and would be a great place for more data centers and outreach centers for academic research.
LANL being the AI version of the Manhattan project is not a serious idea. First of all LANL does not have anywhere near the kinds of people we had in the 40s, 50s, 60s, 70s, 80s and 90s. 2nd LANL no longer has the kind of intellectual culture or culture of excellence it had back 40 years ago. 3rd there is a lot of people that would very resistant to trying to bringing a culture of excellence back to LANL. If you listen to the AI push at LANL it comes across
as very superficial and on the public-hype hairdresser level and not at all serious. If you know anyone in the actual field or AI, they already have much more advanced plans, new directions, new hardwares, combinations of neuromorphic, and new leaning approach. LANL just wants to do ChatGP because that is what is in the news.
A more cynical view is that this is just about getting more money to LANL and they already know we have no real hope of making advances in the field or that the implimentation of standard AI methods will have limited value.
The cynical view is, unfortunately, correct. LANL doesn’t have more than a dozen qualified AI scientists. Yet we probably have twice that number of ill informed managers pushing the technology. It’s weak.
6/01 & 6/02 that’s exactly why the Secretary had this conversation at LLNL as the original post stated….
"Anonymous
The cynical view is, unfortunately, correct. LANL doesn’t have more than a dozen qualified AI scientists. Yet we probably have twice that number of ill informed managers pushing the technology. It’s weak."
The Secretary might want know that LANL had several early efforts in AI/ML 10 to 20 years ago and that LANL management culture did not want to support this non-mission "sandbox science". Almost all those people left and now LANL has a well deserved reputation as a place that will not support real cutting edge and innovative science. We seem to be in a "follow the money" mode and we never innovate anymore or create new fields. 20 years ago it was all nano, followed by additive manufacturing, followed by exoscale.
What makes the AI thing so odd is we have managers who have used ChatGPT to write a memo and think ChatGPT can easily replace scientists. It is truly bizarre.
Post a Comment