Actual post from Dec. 15 from one of the streams. This is a real topic. As far as promoting women and minorities even if their qualifications are not as good as the white male scientists, I am all for it. We need diversity at the lab and if that is what it takes, so be it. Quit your whining. Look around the lab, what do you see? White male geezers. How many African Americans do you see at the lab? Virtually none. LLNL is one of the MOST undiverse places you will see. Face it folks, LLNL is an institution of white male privilege and they don't want to give up their privileged positions. California, a state of majority Hispanics has the "crown jewel" LLNL nestled in the middle of it with very FEW Hispanics at all!
Comments
If AI eliminates 30-50% of the non-productive people state and federal level then it isn't a bubble and the hype is justified. What will happen is people will come to work and just talk, gossip or play on their phones for 30-50% more the their time.
For example at LANL and LLNL there is simply no metric for efficiency however their is metric for spending money and growth. Most metrics, many which have been discussed on this blog have shown the labs are becoming less efficient so the entire culture at the NNSA labs is at odds with the point of AI.
LLNS, has a high resolution employee ranking system, with poor resolution. Thousands of LLNS management people-hours $$$ are spent EVERY year on this process. It would be more efficient and accurate to revert back pure rank groups, and reduce the number of managers, but that won’t happen because the ranking system as is, is too lucrative for the rankers.
LANL is about spending money so (1) cannot be achieved LANL is hiring more people and lots of overheard people at that.
(2) Could in principle work but there is no way measure this. The lab is really measured in how much money was spent. How does AI give you 30 pits a year? Also at seems like people are spending less and less time on work and spend on chasing money, paperwork, forms, training, and fixing the endless messes. No one wants any kind of measure of productivity.
In summary AI may change the way work is done in much the nation but I see it having no effect at the NNSA labs. Sure we may have AI but will add nothing to efficiency it will just be another thing to increase inefficiency.
I am sorry to be cynical but if you ask most people this is what the think.
This was written by AI ;)
Who is writing this junk?
There a many algorithms that have been developed over the years for CNC automation but these are not AI, or least not the AI that the original poster is talking about.
What all the hype is on large language models and foundation models such as chat GPT.
Computer optimization of industrial process and process engineering and has been around for years and years and is not AI.
Until a boss can come up and say "AI make me such and such i need it by end of the week" it aint replacing you at all.
Yeah right. Nobody is seeing 13x productivity gains over traditional algorithms. More like 0% to 25%, best case. Read the studies, they are not extremely compelling for LLMs.
I think the real advancements in CNC programming is either not AI but some other optimization methods or it is some kind of ML but not LLMs.
REPLY
Yeah right.
Just to be clear this thread is about AI, everything mentioned has nothing to do with AI.
Correct! The topic is about AI related lay offs at LLNL, or arguably LANL. Which jobs are most vulnerable…
To be cynical no jobs are vulnerable. LANL and LLNL are not companies trying make a profit. This is the great contradiction of AI at the labs. Part of labs role it to dissipate money "force for good" and that kind of stuff. I see no mechanism for which AI could replace any kind of workers at the labs. Not to mention efficiency is not really the strong point at the labs.
Now in principle AI could make the labs more efficient like like allowing deadlines not to lapse as much or help get to to 30 pits. The question of if AI can replace production facility people in any industry is not clear or if that is possible. To be clear AI has not at this point made any great advancement in industrial jobs. Remember all those AI companies have not even made a profit. Sure there stocks are high based on "future" potential earnings. Many people think AI hype is bubble and the promises are overblown, who knows.
What AI could replace are secretaries, paper work, bureaucracies, data entry, or data searches type jobs, call in centers, services and so on. These of course are millions of jobs like this in the US and thousands of jobs at the labs like this. A company could save lots of money by making HR all AI, having half as many admins for instance. However there is no way on earth this would ever happen at LLNL or LANL.
To more clear I think AI will lead to more jobs at the labs and more bureaucracy.
LANL and LLNL are both laboratories (places), LANL in NM is managed by Triad National Security, a non-profit LLC, and LLNL in CA is managed by LLNS, a for-profit LLC.
Exactly the efficiencies that that AI would bring companies and lead to more product and more profit simply have no meaning at the NNSA labs.
We have huge overheard rates, beyond enormous, and in principle AI could save some of the costs from overheard, so you could do used the saved money to pay for more tech and science staff, better equipment, new buildings, bigger computers and so on which could lead to better efficiencies for the labs but again part of the labs mission is to spend money and employ people. LANL even has this as one of its institutional goals (force for good). Politicians trying get money for the labs are doing so to employ more people they are not trying to get more results from the labs they just want money, jobs and they want the to be spent on as many people was possible. Under this approach why would you want AI to replace people?
In fact if AI does replace people in the private workforce government funded places like labs will be under more pressure to grow and hire even more people.
Bro, don't worry it's just like what happened with the South Sea bubble, this time more spread out over several companies.
Nothing like "we deserve more money because look how much money our stuff is worth and our stuff is the shares that keep going up in value because they're worth so much"