FistLast2
20d ago
As someone who has worked with quantum algorithms for quantum chemistry, that's. . . really silly, and kind of not turning out to be a practical or useful way of approaching problems, despite what Richard Feynman might have thought before quantum computers were really a thing. Addirionally, any claims that quantum computers have a low enough noise floor or long enough coherence times to do any useful calculations are currently overhyped, misleading BS, and it's not clear that there's a clear path out of that.
FistLast2
20d ago
My response to that would be tongo back to something like chemistry, which is computationally more complex than chess, less complicated than modeling a lot of other real world things, but also obeys known equations.
There are some surprisingly simple systems for which all our normal computational chemistry approximations fail and they require much more sophisticated solutions. And you can't always handwave it away to "AI will find a simpler approximation that works". How do you know? Is there a good enough approximation that "works" for factoring any large number? Why should computational scaling laws cease to apply in theory? Would, for example AI be able to solve any arbitrary NP hard problem even if we could prove P != NP?
FistLast2
20d ago
AI doesn't change the problem of computational tractability. There are a lot of problems now where we know how to find the exact solution (to any arbitrary, finite precision) but where the solution is many orders of magnitude beyond what computers would be able to achieve in any reasonable timeframe, even assuming Moore's law. Like the exact energy spectrum of any medium sized atom in a reasonable basis set (yes, there are various approximations that can be computed which work well enough) because the problem scales factorially with the number of electrons and basis functions. There’s so much handwaving in "omniscient" where people are glossing over any serious thought about what it would actually take to achieve omniscience, and at least some of Yudkowski's arguments about the way a superintelligence could infer physics from between 1 and 3 frames of video are provably wrong if you know anything about math and physics (I wrote something about this on an SSC thread perhaps 8ish years ago).
FistLast2
22d ago
About 3 years ago i applied for a government job where i had to take a standardized Pearson VUE test. It covered some algorithms, SQL-like questions, calculus, and statistics. I didn't take the job in the end, but they have definitely been using standardized tests for a long time.
FistLast2
1mo ago
Well, Ive got to pick up dinner, so i can't spend much more time on this, but I'll send a quick response befor I head out.
First of all, I had assumed you quoted everything you consider relevant to this discussion. Clearly that wasn't the case, so I'll look at the additional context you just provided.
For Q1, I'll have to watch the first 40 minutes at some point to make an informed judgment here. I can think of many reasons why this wouldn't be unreasonable--for example, the negotiated deal contains no security guarantees to Ukrain and no promise of support. My assumption is that Zelensky was hoping it would be an in to negotiate further support. It's possible that he did this in an unreasonable way--granted I haven’t personally watched it so I cant say for sure--but in itself, this is how normal negotiation works. He agreed to something that doesn’t really benefit Ukraine with hopes of getting something more in the future. And again, it's certainly possible that he did this in a blatantly offensive and unreasonable way. Based on what I've seen, I doubt it, but I'd have to watch it myself.
Q2: Why on earth would you do this? It seemed blatently apparent to me from the course of the conversation. I will reiterate my take on it that I already posted:
Vance: Negotiate. Zelensky: We tried and it didn't work. What do you want us to do? Vance: Negotiate for peace, loser. Zelensky: But Russia brome our peace deal last time Vance: How dare you talk to us lime that?
I mean, if your take is different, fine, but don't just expect me to take it as a given that I'm going to agree that what appears to me as the obviosuly correct interpretation of events is wrong.
Q3: Again, I'm not seeing it. I know this is your take and apparently Hanania's, but as I keep saying, it does not look like that's what happened to me at all.
Anyway, I hate typing on my.phone and I have to get going.
FistLast2
1mo ago
I don't really see what this adds to the discussion. I mean, if we start from the framing that Zelensky suddenly became hostile out of nowhere, then, yeah, that would make Trump and Vance upset. My entire contention has been that that's very clearly not what happened, though. When I tried to point out that Zelensky didn't appear remotely hostile to me, you said something about this whole situation being caused American intelligence fabricating evidence against Trump. And when I effectively asked "OK, but what part of what Zelensky did in the meeting was actually hostile," you posted someone stating that Zelensky was hostile. Like, sure, I know you think that already, but you still haven't shown why saying Putin is untrustworthy should be interpreted that way by the US. I know that was the mo.ent JD Vance chose to put on a (very obvious, IMO) show for the cameras, but it’s kind of absurd.
Let's take everything you're saying at face value and I'll concede what I'd otherwise consider dubious or false claims in your post.
What do you think the point of even inviting Zelensky was? What concessions was the US expecting that he didn’t offer? What SHOULD he have said in that meeting to make things go more smoothly? What part of that indicated that Zelensky was, in the moment, changing the terms of any deal the US and Ukraine had planned to agree to? All he did was point out that Russia can't be trusted to hold to the terms of a peace deal and that Russia will not be a good ally to the US long term. He was calm and defferential in the dace of a lot of insults. The things you just talked about have nothing to do with what Zelensky said or did in that meeting. He said thank you over and over again and didn't argue with any offensive comments made toward him peraonally. All he did was point out that Russia isn't a trustworthy ally, which aside from Trump, the US generally acknowledges. So what did you expect Zelensky to do differently TODAY and why?
FistLast2
1mo ago
Can you find the part of the deal where Trump offers to bail out Ukrain in any way?
I'm really confused about how anyone is interpreting this as Zelensky being unreasonable or childish. Trump spent the first 30 seconds essentiay badmouthing Zelensky, during which time Zelensky sat there quietly and didn’t react. Around the 1:20 mark when Vance mentioned negotiation, Zelensky very calmly and defferentially said "Can I ask you something?" It wasn't in a heated or charged way--he was clearly trying to engage in rational discussion. He even said, "I'm not speaking just of Biden, it was president Obama, and President Trump, and President Biden, and now president Trump, and God bless now President Trump will stop him. But in 2014 nobody stopped him (Putin) he just occupied amlnd took. He killed people. . ." And he goes on to piint out that for 8 years, he tried to negotiate with Putin and that he actually signed a ceasefire in 2019.
What part of that is problematic? What part of that should have caused negotiations to blow up? You're telling me that somehow Zelensky didn't phrase something exactly right and that set Trump off and Zelensky's at fault because he shojld have known better? Not the guy who flies off the handle when someone says something that in any way challenges what he wants to hear, no matter how calmly, kindly or rationally?
Everything from Vance looked to me like a performance intended to make it appear as if for some reason Zelensky was doing something inappropriate, but i can't figure out what he's actually complaining about. "We tried to negotiate. What do you want us to do? ", "you didn't do negotiations that would bring peace, dummy. How dare you come here and say otherwise in front of the media?"
I literally can't believe that anyone is falling for Vance doing anything other than trying to put on a self aggrandizing performance here.
Nobody's immediate supervisor is telling anyone to ignore the email. The directives are coming from agency heads or higher--Trump appointees like Kash Patel and Tulsi Gabbard. People in my agency didn't even get the email, although the two Trump appointees who are in my chain of command have told their people not respond if we do.
Edit: I should say nobody's immediate supervisor is using their own discretion to tell people not to respond. The information is passed down from higher up through the chain of command.
FistLast2
1mo ago
I recently left pricate industry to go work in the government (DoD specifically) for the second time in what is legitimately the most important job I've ever had. I've always worked in research and tech, so government is a big pay cut, but I also really like what I'm doing, and I like that work doesn’t take over my life even though I'm legitimately contributing a lot.
My issue with this whole thing is this:
First, everything I work on is classified. Sure, I could come up with some unclassified bullet points like "developed Python codes", "worked through mathematical problems", "investigated some API's, fou d other data sources", "worked through more complicated math when the first approximation wasn't good enough", but what is this proving? How does it show my job is necessary or valuable?
Second, my boss, my bosses boss, etc. up my chain of command know what I'm working on and they're all impressed. Are my five bullet points going to be used to somehow override theor judgment of me? We al we already have to do evaluations at work. We do self evaluations and our managers evaluate us. My evaluations are really high. They always have been in every job I've had. Why the fuck do I have to send 5 bullet points to Elon Musk and be subject to some randos judgment based on that?
Finally, I read a news article that said AI is going to be used to determine which jobs are necessary based on these bullet points. I'm not sure what the source is on that--maybe it's wrong. It seems likely to me, though, because there's no way they're going through each email by hand. Again, fuck that. My job is legitimately important, for reasons I can't talk about here.
For all the complaints I see here about federal beurocracy and waste and the cold, methodical nature of beurocrats, I have to wonder, what is thay based on? Federal offices are just, like, normal work places. And the people who work there are normal people, who would be very unlikely to enact some conspiracy of hiring fake employees, faking vasge swipes, faking timecards, etc--no more likely than anyone else, certainly.
One more comment: In my previous industry job, we got a new CIO at one point. He really liked my boss, who he promoted to a position directly below him, and she knew how valuable I was, and so the CIO ended up really liking me as well. He put me in charge of some stuff, and I talked to him semi-regularly.
The problem was, he had no idea what he was doing. He laid off all our middle management in tech, which a year later we still hadn't recovered from. He snatched up tons of people who were laid off from Amazon, Google, etc. bloating our AI/ML. He decided he wanted us to move off the cloud and develop tech that we could sell to other companies, even though it wasn't a tech company.
He started canceling contracts with vendors. Firing contractors. Building on prem systems. I had VP's and their assistants screaming at me that we were blowing things up and I had to change the CIO's mind. . . But of course, i couldn't do anything. High up people started quitting.
At one point (this wasn't actually his fault), I was trying to requisitions funds for a project I'd veen told to start, which involved keeping 3 contractors who did very important work on our financial syatems. I argued and fought but finance kept telling me we couldn't keep the contractors, even though they were the only thing preventing another group of external contractors from pushing changes that would routinely destroy our payroll systems.
Anyway, long story short, I quit that job and came back to the federal government. I'll be damned though if it isn't starting to feel like a repeat of the same situation.
FistLast2
2mo ago
Normally? This is at best a Chinese robber fallacy and at worst a series of cherry picked, misrepresented, and false claims. That is absolutely not the way it "normally" works for federal government employees (not that I don't understand what you're referencing).
Where are you getting that? Ive never seen a classified cafeteria menu. Part of your classification training specifically warns about overclassificstion. It got hammered into us quite a bit that it's just as bad as underclassification. Aside from the administrative issues it causes, people have actually been killed because information was classified above the level that the people who needed it could get at.
Edit: and although there's a process for declassification, it's long and complicated by design, because classified information is supposed to be handled with care. This is probably also part of the reason people are freaking a out about Musk and USAID classified info--there are normally extremely severe penalties to anyone who hands classified information to someone who doensn't both 1) have the right clearance and 2) a need to know. The idea that this process is being circumvented so easily when normally in situations of much greater consequence the rules normally hold fast is troubling.
I've been reading here since the SSC days but only posted a short message once before. I wanted to weigh in on some of the stuff discussed in this thread, though.
I'm currently a federal employee for the second time, and evem though I'm at the top of the GS scale with the DC cost of living adjustmeny, I still toom roughly a 100k pay cut to come to my current job. Largely, it was because of stress and hours worked, and because my wife, who works here, liked it so much, although with the commute (I had to come in 5 days a week prior to the EO but worked from home in my private industry job), it ends up being a long day anyway. The first time I was a federal employee, I was a theoretical physicist at a DoD lab--unlike DoE national labs, where the employees are contractors, DoD lab employees are federal employees--but I made a lot less money than I do now.
First of all, I wanted to mention that only one person I know where I work has received this buyout email. I didn’t. If I had, and I thought it was legitimate, I might take it, because I wouldn't have trouble getting another job. I somehow still end up getting emails from recruiters despite having deactivated my LinkedIn account. I have literally no reason to be here if they’re going to try to turn it into a bad work environment.
Second of all, with respect to DEI--the office that our agency had previously labeled DEI mainly handled EEO complaints and reasonable accommodations. DEI was almost like an afterthought and essentially only ge erated various legally mandated demographic reporting requirements and planned the occasional heritage celebration event, whoch was always optional. Both EEO and reasonable accommodations are still mandated by law, so they can't get rid of those--so now we just have an EEO office which is mostly the same as it was before, except maybe now they don't have to generate as many reports (although that's not even clear). They've also done some silly things like remove letters from various acronyms if they could be construed as relating to DEI, so the acronyms don't actually spell words anymore.
The general sense I get when I talk to people at other agencies about the "buyout" is that there's a lack of belief in it's legitimacy. If you reply "resign" to an email, is there any gaurantee that they have to pay you for the full 8 months? Can they fire you or lay you off in that time? Can they require you to come into work?
Finally, anybody who works with money in the federal government knows all kinds of ways that government agencies could save money. Not rewarding people for spending their entire budget and penalizing them for having money left over for example. However, the executive branch is powerless to change most of this. Laws, congressional oversight, and demands for transparency and congressional control force things to get done in a particular way. Frankly, whatever small amount of money gets saved by trying to get workers to resign and through other reductions in force (which i expect won't be as significant as some people here think. The hiring freeze, for example, doensn't even apply to the DoD) will be a drop in the bucket compared to what could be done if things were done in a smarter way. Is DOGE going to somehow make it possible to do those things? I'll believe it when I see it.
Edit: I almost forgot, regarding the J1 J2 thing. Maybe that's illegal if you're doing them in the same hours, but my agency explicitly allows you to have outside employment if you get it approved and if it doesn't interfere with your work schedule. During my old non-government job, I was working weekends for a startup, and I considered trying to get it approved so I could continue here, but ultimately, I didn't want to spend the extra time working.
At the end of the day, that's not really what matters, because nobody is going to need to solve a problem in physics with a known solution. A good portion of tests that I had as an undergraduate and in graduate school were open book, because simply knowing a formula or being able to look up a particular value wasn't sufficient to be able to answer the problem. If I want a value from NIST, I can look it up. The important part is being able to correctly engage in the type of problem solving needed to answer questions that haven’t ever been answered before.
I've had some thoughts about what it actually means to be able to do "research level" physics, which I'm still convinced no LLM can actually do. I've thought about posing a question as a top level post, but I'm not really an active enough user of this forum to do that amd don't want to become one.
Finally, I want to say that for the past 18 months, I've continually been getting solicitations on LinkedIn to solve physics problems to teain LLM's. The rate they offer isn't close to enough to make it worth it for me, even if I had any interest, but it would probably seem great to a grad student. I wouldn't be surprised if these models have been trained on more specific problems than we realize.
More options
Context Copy link