"Top engineering school" probably makes a big difference. I too went to a top STEM school (Caltech).
Most STEM classes were taught at a fast pace, and later classes assumed you knew the material from your early classes well.
Occasional cheating might work but anything more would almost certainly send you on a path of falling farther and farther behind, requiring more and more cheating, and there were enough things where you would have to work with other students on projects that people would quickly find out you have no idea what you are doing.
Hundreds of thousands of tech workers have been laid off in the last two years. There are many skilled Americans on the job market. Foreign workers are here for the benefit of the United States, not as charity to the rest of the world, or as a source of cheaper labor for global employers seeking to undercut American wages.
It's clear you are not a hiring manager. There aren't thousands of highly qualified engineers looking for work. I can't tell if this a troll post or not.
It's clear you are not considering the long-term effects of labor policy. If we want a stronger domestic workforce, the answer is to incentivize Americans to develop the skills we need and to train them when they do not already have those skills. Mass importing foreign labor is not a substitute for building and investing in the domestic workforce.
Btw, this is analogous to the argument for positive discrimination:
Preferentially treat one group of people (non-white/women/Americans) even if in the short term the revealed preference of businesses is to hire a different group of people (white/men/immigrants), such that, in the long term, the first group of people can develop the necessary skills to compete with the second group.
The arguments can be the same but have entirely different hidden variables that make it fine to use the argument in one context and not the other.
Also, this ignores that the government works for citizens, no matter their race. Promoting racial equality among citizens is a function that the government might undertake. Pursuing policies that favor citizens over non-citizens falls in line with that same reasoning.
No it's really not the same because we have laws against racial discrimination, but it is the entire purpose of the state to represent the interests of its citizens above all others.
You can’t get them to apply. They are already very well paid. It’s a finite pool that is not large enough. Bidding up pay won’t change that. Instead you just hire overseas.
working with what tech? you didn't answer the question to a reasonable extent. it seems odd you need to hire people into principal and "distinguished" titles, you should be promoting from senior/staff to fill those internally. unicorn hunting is an organizational problem, not a useful reflection of the hiring market
Fortunately such ridiculous hypotheticals have no correspondence to reality. Most every job can be easily trained or cross trained. Mathematicians make good physicists. Nor is there some foreign pipeline of a million physicists. Such ridiculous argument exist only to perpetuate the lie of a skills shortage and deny the wage shortage.
They may make good physicists, but not great physicists. Similar to saying I know how to drive a car, now let me win the inde 500. Things take to for specialization, even with the advent of guessing and checking machines (llms).
A shocking number of innovations in x field are because y person brought an idea that maps from z field. Cross functional teams are important and mathematicians can and absolutely do make top level contributions to physics.
One of the math problems recently publicized as having being solved by was actually just discovering that a German mathematician had published a solution in the 1950s.
Oh that is true, and I’m not putting down cross functional teams. I’m just putting down the idea that a specialist in one area can immediately be a specialist in another semi related area with no extra review time.
If you ask specialists about their backgrounds you will find that they took different paths. The critical step is time in the speciality, not background. For any job there's never a pool of applicants who are pretrained for that exact job, everyone takes time to get up to speed. People from related fields do just as well learning job tasks. The labor market is fluid if properly incentivized. Commenters intentionally pretend not to understand because they want cheap labor and want to impoverish the American worker.
Commenter doesn’t seem to understand how much of a drift there is for specialty learning or different negative patterns specialists may have. A front end architect and a computer architect may have similar backgrounds, but a computer architect will take wildly longer to learn front end and likely have some practices to un-learn.
Also I would say for certain areas, people are pre-trained for the exact job as the tools or tech between is the exact same.
It’s not just commenters who want cheaper labor, capitalism wants cheaper labor.
A lot of employers would have slaves if they were allowed, doesnt mean that the workers shouldn't do everything possible to prevent this and politically demand better conditions. Such as the case in the topic of the thread.
I would agree employers would likely want that, though since in the us it’s difficult for labor unions to ever be created let alone extra protections or the necessity of tips or us at-will hire/fire.
My point of your thread was disagreement with the immediate transfer of knowledge: “Most every job can be easily trained or cross trained.”
I disagree with most jobs can easily be trained or cross trained, there is more of a time sync between that.
The national government functions as a labor union through its control of the border and regulating the conditions of employment.
>I disagree with most jobs can easily be trained or cross trained
People just overestimate how complicated their job is because it makes them feel important. If anything immigrants take longer to get up to speed which is why companies usually only hire them when its cheaper. Or nepotism.
“The national government functions as a labor union through its control of the border and regulating the conditions of employment.”
Woah for some reason it sounds like you’re talking about a European country, not the US where the legal age to work changes every few hundred miles. 1
“People just overestimate how complicated their job is because it makes them feel important.”
I do hate to say but I also really doubt this anecdote for majority of workers. Wasn’t there a ford factory that laid off workers in favor of llm (that were trained to replace them) only then to be rehired? 2
So we’re in agreement? The time of investment for growth is the point of decision, and how long it takes to train specialists defines if it’s worth it.
I.e. how long would it take to train a regular driver to drive the inde 500 vs buying a racers contract
Want me to agree that training takes time? Ok, but what a boring revelation. It doesn’t take much time in the grand scheme of things. I can read a substantial book each week, for example. Give a person a half ~salary and let them study and train for a month +/-, they’ll be ready to contribute.
Good thing every single requirement is known in advance. It’d be a shame if things changed faster then a snail, limitations we’re fully understood, and the geopolitical situation doesn’t change by a tweet.
Edit: actually, what job do you have in mind that never changes? Honestly it sounds relaxing
The current admin will be gone or dead soon. “Beef prices” are not a factor in the tech industry.
Just a generation or two ago we went to the Moon. Now, you refuse to train an intelligent person with overlapping skills.
Not just a closed mindset, but pure defeatism.
Why even get out of bed in the morning, right? Might as well give up now.
Back to reality—I have an acquaintance who started work at JPL and finished her degree while working there. But according to you it’s just not able to be planned for! Thankfully defeatists are not in charge in Pasadena.
Good thing people aren’t affected by beef, by eating it, getting diseased from it, or from sick from other people eating it. And nice choice picking 1/3 of the options.
This is not me refusing to train someone, but acknowledging there is a time sink cost for training that may not pay off in the need to make a project, or the person could up and leave which good for them.
Thank you for accelerating my argument to 1000 when I was just going the speed limit and seeing the warning signs for the problems up ahead.
Yes, you repeatedly tried to refuse training as a possibility. My initial statement was relatively conservative so can’t be easily disproven, but you made many attempts anyway.
Others discounted the exaggerated emergency above so no need to repeat. And we know training worked for centuries up to my example of two years ago. Most people don’t study exactly what they end up doing and learn as they go. The US has a couple hundred million workers. All which point to training being more economically efficient than importing from abroad except in extreme cases.
Your closed-mindset, anti-worker ideas and position continue to harm the economy and wages. Enjoy the fruits of your war on the middle class, having to live in a luxury bunker.
(Oh and gas and steel prices don’t materially affect tech either.)
I can't tell if this is a troll post or intentionally obtuse. The comment they you replied to did say that these are the best in their country, which is why we want them. Have you heard of a concept called brain drain?
Yes, I have heard of brain drain. And people here are arguing in favour of brain drain rather than hiring citizens or paying citizens better, because they say that US needs these foreign workers. That's why I ask how it comes that their own countries do not need them?
Why should anybody be sympathetic to the so called "needs" of US companies hiring immigrants, but not be sympathetic to the needs of their home countries?
Sympathy for companies is usually conditioned on the fact that they give jobs to locals, or at least to citizens. If they're not doing that, then who gives a fuck if they aren't able to expand or profit at a rate that shareholders prefer?
But there is a class of people - most hackers included - who have had such comfortable lives that they aren't capable to even pretend anymore to sympathize with the producing class.
I’m very pro immigration, but “Americans are lazy and stupid , thus unfit for hiring” isn’t a winning argument.
Tons of Americans are out of work. I’ve personally seen my wages drop by at least 20$ an hour as well as the work being less stable.
Gasp, you might need to train someone a bit or have employees that will say no if you ask them to work on Christmas.
If I ran the H1B program it would be an unrestricted work permit with an automatic pathway to a GC after 5 years. Don’t get *convicted* of a felony and here you go.
> If I ran the H1B program it would be an unrestricted work permit with an automatic pathway to a GC after 5 years. Don’t get convicted of a felony and here you go.
H1Bs create new companies here, we continue to attract the smartest people here.
However each company would have to weigh the risks of bringing someone to the US and they then walking away.
We could even change it so instead of a 100k fee, you have to invest 200k in US bonds which only mature after 5 years. Of course even if it doesn’t work out you still get your money back.
1) It benefits an extremely tiny slice of American entrepreneurs by allowing them to operate directly against the best interests of the broader American labor force?
2) This posits that there must be a large pool of highly trained and specialized Staff Eng sitting idly in other countries just waiting to be poached. What are those countries doing which is generating a huge untapped pool of labor? If that _is_ the case, then why aren't entrepreneurs in those countries capitalizing on the idle labor?
---
Maybe I'm entirely too cynical - but I simply don't believe that #2 is true. There isn't a huge Indian talent pipeline just pumping out hoards of incredible world-class engineers worthy of poaching, yet America is entirely incapable of producing the same tier of engineer. I think it's purely a function of depressing labor costs at the lower end of the labor ladder and over the last ten years has actually just been an immigration pipeline with pay-to-play schemes on both ends.
Hiring is such a shit show, there's no reason to assume the people doing the hiring know what they're doing. Like HR has to insert them selves into the process
> Foreign workers are here for the benefit of the United States
But that's the point, isn't it? Some foreign workers who "worked the US for the benefit of the US" are now unable to come back to the US. Isn't that bad for the US too?
And what do you do about the people who only want to work in jobs that they're not qualified for, and who have unrealistic expectations about pay, benefits, work demands, etc.?
Is every natural-born American entitled to a job that pays what they want, for the hours they want, in the location they want, regardless of their qualifications for the job?
But isn't that how market supposed to work? Company offers to pay x for the job, can't find anyone for x, so it raises salary to x+y.
I would understand logic of bringing people from other countries when there is a shortage of workers but with current market it doesn't seem to be the case.
Of course not, but you seem to miss the point: no company is going to pay $500,000/year for a part-time office assistant, $200,000/year for a bus boy, $2 million/year for a construction worker, $100,000/year for a farm laborer.
There is no developed country on earth where labor shortages are addressed by "just increase the salary until you find a native-born person to take the job."
This is a misnomer is _my_ point. A labor shortage has a trivial solution, "pay more". It's intentionally framed as "labor shortage" by capital who would prefer the labor pool be near-infinite so they can pick the best for the cheapest. No, no company is going to like it. And that's okay, immigration, and government policies more broadly, should not be about "what would maximize private capital profits?" They should serve the broader American public.
If companies can't find a farm laborer at 100k/yr then that sounds like it's an *incredible* situation for the average worker.
> A labor shortage has a trivial solution, "pay more".
Whenever you think a solution to a problem that plagues an entire nation (especially one the size of the US) is "trivial", you should ask the question: if it's so trivial, why hasn't it been fixed after so many years? And if your answer depends on some sort of conspiracy, you know you're wrong.
There are so many problems with "pay more". Let's start with a simple one: it's based on the assumption that the employer's only alternative to paying more is going without the output. In a lot of labor markets, this just isn't the case.
Jobs get mechanized/automated. They get outsourced to a different country. The employer reduces output (reduces staff) and in many cases at the same time increases prices. Sometimes, as has been seen with restaurants and retail, hours get cut or they go out of business.
Another problem with your "solution" is that in many markets, especially those related to healthcare, it's not just pay that limits hiring. It's things like caps on residency slots due to Medicare funding, a lack of clinical placements, a lack of faculty, etc. You can't pay more to hire workers that don't exist because the market structure doesn't allow for those workers to exist in the needed numbers.
> If companies can't find a farm laborer at 100k/yr then that sounds like it's an incredible situation for the average worker.
Except this experiment has been tried. Look up the Bracero program, which was terminated in the 1960s to increase wages for domestic farmworkers. I'll give you a hint: a study comparing states heavily exposed to bracero labor against states that used little found basically no wage or employment gain for domestic farmworkers.
The farm doesn't survive with a bunch of recent university grads who studied CompSci picking fruit at $100,000/year. The farm mechanizes, where possible, and switches to crops that require less labor. And the market fills the demand for the crops that were switched out with imports.
And let's also be honest: the average 20-something native-born American doesn't want to pick fruit in 100+ degree heat in dusty small farming towns where there's nothing to do, the male-to-female ratios are out of whack, etc. This too has been tried: UFW ran a "Take Our Jobs" campaign about a decade ago and...wait for it...literally nobody wanted to.
Migrants don't really depress wages, except by a very marginal amount at the lowest skilled rung.
That's because adding people to the economy creates more demand
In times of recession, things might be different - but immigration tends to take care of itself in that regard. When a country has unfilled labor demand, immigration goes up. When it has no labor demand, immigration goes down.
I was homeless for 6 months, and wanted nothing else except for a job that can pay for basics. I still don't have dental, can't save any money, but I have a room with a lock. I worked from 5pm till 4:30 in the morning from covid until a year ago with zero benefits!
Will I ever be able to retire, doubtful. Have a family, probably not. Are these personal failings I need to take responsibility for, absolutely. Still my expectations are very low and my earnings and lifestyle reflect this.
Where do you want me to move? What job do you want me to take? What more do you want to take from me? Do I deserve less, more, just enough? Why do I have to listen to you about that?
> I was homeless for 6 months, and wanted nothing else except for a job that can pay for basics.
That's entirely reasonable but for you or anyone else to get such a job, two things are required: you need to have the capability (knowledge/experience/skills) to fill an open role and a prospective employer needs to offer you compensation that you're willing to accept for filling it.
> Where do you want me to move? What job do you want me to take?
I don't want you to do anything. If you want a job, don't you have to figure out what you can do and where you can go to find one? I'm not going to suggest it's always easy, because it's not, but if you can't find a job that can pay a wage that meets your needs and expectations, don't you have to look at what you can offer and/or where you can go to find such a job?
What is the alternative?
> Do I deserve less, more, just enough?
How would I be able to answer that? What you actually get in today's market is a function of many things, including the skills and experience you have, where you live, the demand for what you can offer versus supply, etc.
> What more do you want to take from me?
Please explain how I have taken anything from you.
> Why do I have to listen to you about that?
You don't have to listen to me about anything. I'm not a prospective employer, and I'm not a politician creating immigration policy.
> You have _no_ clue what you are talking about.
With all due respect, you don't know anything about the personal situation of anyone you interact with online. You started your reply out with a discussion of your difficult experience, and while I empathize with your experience and think it's indicative of America's problematic economy, it doesn't make you uniquely qualified to tell other people that they don't know anything.
Demographics are real. Immigration is a given in developed countries with aging populations. The idea that there's ever going to be a way to guarantee every native-born American (or Canadian, or European, etc.) a job before an immigrant gets a job is just fantasy.
There are many problems in today's labor markets. Youth unemployment. Underemployment. Wages that haven't kept up with inflation. The cost of living crisis. But salaries just aren't going to grow without limit until employers find a native-born American to do the job. It doesn't work that way and it never has.
But America is a capitalist system. The market has determined that not all Americans deserve jobs. Are Americans willing to work a 996 schedule for 2.00/hr to compete with labor overseas?
If your business cannot survive paying more than slave labor rate, then a capitalist system dictates that your business should cease to exist or your resources allocated to more productive work.
Capitalism is what capitalists do. No capitalist system dictates that, because that no capitalist system practices that. Wages don't increase with productivity anymore, the gap between the two has only increased since the 1970s.
Capitalists come out of the woodwork to denounce things like the minimum wage or unions that distort the fair market value of labor because they want to pay as little as possible and not a cent more. Why do you think automation and outsourcing exist to begin with? Why are we putting Manhattan Project scale effort into AI? Wages are an inefficiency in the market that capitalism has always sought to ruthlessly optimize away.
Most capitalist countries have 100% employment for their citizens and little to no immigration? No. "Every citizen who can work having a job" is not something capitalist countries do. They all take ample advantage of foreign labor, they all have corporations that prioritize their own profits and outsource.
> Most capitalist countries have 100% employment for their citizens and little to no immigration
You must have responded to the wrong post as it has nothing to do with my statement.
That statement is that there is nothing about capitalism that prevents a country from prioritizing its own citizens. That is a perfectly normal and rational thing to do.
Additionally, the exact percentages of unemployement rates also has little to do with capitalism, nor does the uncommonness of fully achieving a lofty but possibly valuable goal also has nothing to do with what capitalism is, nor does it say very much about how valuable that goal is or the value in attempting to aspirationally achieve it.
>You must have responded to the wrong post as it has nothing to do with my statement.
OK. You can't handle context, noted.
>That statement is that there is nothing about capitalism that prevents a country from prioritizing its own citizens.
The free market is what prevents it. The purpose of capitalism is the creation of capital and its transfer to the capitalist class, and the purpose of corporations is to maximize profits to facilitate that process as quickly and efficiently as possible. Neither capitalism as a system nor corporations have an obligation to improve quality of life for their employees beyond their market value, or to "prioritize" the social welfare over profit, or to hire more people than they need.
It's simple supply and demand and there isn't a demand for American labor.
>Additionally, the exact percentages of unemployement rates also has little to do with capitalism
which is weird because you're arguing that capitalist companies need to prioritize "citizen" labor in order to reduce unemployment, or that governments need to force them to.
> Neither capitalism as a system nor corporations have an obligation to improve quality of life for their employees beyond their market value
If by capitalism as a system, we are refering to countries that are capitalists, it seems fairly uncontroversial to say that such countries can prioritize their own citizens.
It is not weird or controversial to say that countries, even capitalist ones, prioritize their citizens in certain cases, nor is it weird to suggest that they aught do so.
> which is weird because you're arguing that capitalist companies need to prioritize "citizen" labor in order to reduce unemployment, or that governments need to force them to.
I am suggesting that it is compatible with being a capitalist country, and is not weird, unusually, or bad to do such a common thing as prioritizing your own citizens.
I wouldn't say that they need to. Merely that it is a normal and compatible, and in many cases desirable goal for said citizens.
Capitalism is at odds with protectionism. Putting citizens above profits doesn't seem like a very capitalist thing to do. Not arguing against it, I'm all for socialism
> Putting citizens above profits doesn't seem like a very capitalist thing to do
Do you... think that people who are pro capitalism believe that their own ideology is harmful? Obviously proponents of a belief think that their ideas and philosophy help their own citizens.
> I'm all for socialism
Thats only because you have definitionally defined socialism as "When you do things that are good", not realizing that other people think that their ideas are also good.
The reason why people support capitalism is specifically because they think it is a good idea and benefits a country, and its citizens.
Do you propose to restrict the international flow of capital too while you're at it? Until every spare US dollar is invested in the US? or you just want to restrict the market for the 99% of people whose biggest asset is their labor.
Thats the amazing part of a diverse society, there is a variety of opinions and world views that you can come across that that might even be completely outside your current understanding. As by being curious and trying to understand something surprising you might just gain a new and deeper understanding of the world.
Most folks don’t want their minds opened to new and surprising views and opinions. They want a good job in the town they grew up in so they can take care of their family and raise their kids in the culture they were raised in.
Globalism has ruined that entirely, and there’s no amount of implying these people are closed minded or incurious that will make them change their mind on this issue.
People were assured that COVID vaccine mRNA would disappear within days and the spike protein would short-lived. Yet Yale-led researchers later detected circulating S1 and full-length spike as late as 709 days after vaccination.
They also left the public in the dark about the risk of vaccine-associated myocarditis, especially in young males, many of whom experienced permanent heart tissue damage.
That is a massive gap between what the public was told and what later evidence found. If a technology makes your cells produce a biologically active protein, regulators should establish where it goes, how long it can persist, and what prolonged exposure does before mass deployment.
> People were assured that COVID vaccine mRNA would disappear within days and the spike protein would short-lived
which is still the case, even in the Yale study.
> That is a massive gap between what the public was told and what later evidence found.
there is an even larger gap between basic epidemiology and what the public understands.
Hard science is difficult to communicate on a good day, on a global pandemic with quacks getting tons of airtime and multiple labs finding conflicting results (due to speed to publish and bad experimental parameters) I would say that was communicated by health experts was surprisingly valid and accurate.
+1 the public health apparatuses' actual statements were impressively correct even if you don't account for the underlying uncertainty at the time.
Where public health officials "failed" was not having a megaphone loud enough to drown out the game of telephone that most people (as these comment threads consistently show) were actually hearing information through.
Simplest example: People will say that public health officials told them "masks don't work." But no! Only one person did, and that was USSG Jerome Adams (who wasn't credible nor a major player to begin with). What public health officials said was "we do not know if masks work," which is a more nuanced but completely accurate statement.
> : People will say that public health officials told them "masks don't work." But no! Only one person did
I think Fauci might have had a statement early.
But tbh every other scientist said the opposite, every doctor said the opposite.
If I am not wrong he later admitted that he said it to gain time for the goverment to adquire enough for basic needs like hospitals.
And that is the kind of error that erodes public trust and it was bad advice shared because the american public cannot be trusted to prioritise medical professionals being stocked and because the american gov was unprepared.
But one bad statement, in march of 2020, done for the wrong reasons due to an unprepared goverment voted in by selfish people who cannot be trusted to help others does not deny that the UK had great advice the entire time, so did China, and Spain and Italy and most of the other initial hotbeds. They said "we dont have enough n95 for everyone, please social distance until they are availeble" instead of "masks are not needed".
Despite that, he was the face of a pandemic plan in a goverment alergic to facts and led by a dude who recommended injecting bleach so in that situation, he did as much as you can and is still paying for it.
Could you please stop posting unsubstantive comments and flamebait? You've unfortunately been doing it repeatedly. It's not what this site is for, and destroys what it is for.
Good to know proposing proper safety research, longitudinal studies, and testing of completely novel technologies before they are emcouraged to be administered into the populace of the entire worldwide human population is considered in such a way.
We don't care what the topic is or what your position is; you just have to observe the guidelines. All-caps words/phrases/sentences are specifically banned because they make the comments ragey and hostile, and make your argument seem weaker. A phrase “I don't give a flying fig if you all die from that hideous nasty crap” breaks multiple guidelines and is almost worthy of an instant ban, but I'm hoping you might be persuadable to try harder to use the site as intended. It isn't difficult, and if your points don’t hold up without that kind of aggression, they’re probably not strong points.
Also, please don't try to reframe what is plainly egregiously abusive commenting as reasoned argumentation. That’s obviously game-playing, not honest dialogue. This site is only a place where people like to participate because others make the effort to raise the standards rather than drag them down. Please only comment here if you intend to participate earnestly.
What? The Yale study is where the almost two year figure came from. Are you just making things up? Did you read what I wrote, or anything in the study?
> The Yale study is where the almost two year figure came from
yes, the figure that a small sample of people had long period of protein in them. But most people they looked at had none, which was what they expected.
The average time of it was not contested, the long tail was.
My comment was not even that hard to read, makes me wonder what kind of other information you have misread to come to your conclusions
> They also left the public in the dark about the risk of vaccine-associated myocarditis, especially in young males, many of whom experienced permanent heart tissue damage.
Not true. This was correctly disclosed in the trial results as a super super super tail risk.
Once they realized that the (minuscule) risk could be even further mitigated by simply spacing out the booster schedule, they adjusted that (Feb 2022).
What's more, they found that myocarditis risk was higher in the unvaccinated population after infection. This information didn't get to the so-called "skeptics," who willfully ignored data that didn't fit their conspiracy theory.
I also get a kick out of people only looking for side effects they think of as side effects, since they are what people think of as side effects for completely different types of technologies.
>In men younger than 40 years old, the number of excess myocarditis events per million people was higher after a second dose of mRNA-1273 than after a positive SARS-CoV-2 test (97 [95% CI, 91–99] versus 16 [95% CI, 12–18]).
From your own study you linked. The first comment I ever posted on HN was regarding a similar study where there was no control group whatsoever for unvaccinated people.
I can’t get over the sunk cost reaction to this vaccine. No one who was not immunocompromised should have ever taken it. Period.
And this is just regarding your study.
Remember how the AstraZeneca vaccine (it was called the “Oxford vaccine” before that bad PR) was taken off the market instantly after reports of blood clots.
Until when? At what point? Where exactly? The Astra Zeneca (Oxford) vaccine is relevant because it was taken off the market almost immediately after the effects were found. The technology behind it is utterly irrelevant to the point that I am making.
How am I a conspiracy theorist when I literally pointed out what the study said. Is it not a fact that the Oxford vaccine was taken off the market almost immediately after effects were concretely found.
Let's get your theory straight. You posit a conspiracy where some cabal kept mRNA vaccines on the market despite harms while removing the AZ vaccine. This theory requires a conspiracy. Hence, conspiracy theory.
There’s no reason to assume frontier-level intelligence eventually collapses all the way onto a midrange consumer GPU. In fact, there are quite a few reasons not to assume that (information-theoretic constraints, etc).
I'm pointing out that there's no known information theoretic constraint about the impossibility of frontier AI models being improved to fit/run on a small GPU.
Please do not make up plausible sounding science facts.
Please do not assert I am making a claim I’m not making. Information-theoretic constraints exist. My comment does not require some specific, hard constraint to have been clearly defined, for my point to be valid.
If I were to say you could put a motorcycle in my car’s trunk, it would be perfect valid for me to say there are space constraints that make your idea unlikely. The same is true in this discussion, even though I have not computed the exact dimensions of the motorcycle and my car’s trunk.
Your claim was about frontier intelligence and midrange consumer GPUs. That's a pretty specific constraint.
Sure, there could be some point between a midrange consumer GPU and a pocket calculator where you can't fit enough 'intelligence'. But we really have no idea if the constraint is information theoretic or something completely different. Demonstrating that is the hard part, not finding the exact number of bits.
Talking about motorcycles in car trunks is just lazy false analogy here.
I will not claim a 5070, but there is already evidence in nature that you can get very good general intelligence with an order of magnitude less wattage.
There are constraints of course- training takes way longer.
I wonder if we'll eventually find that Darwin style evolution gets us close to the global optima of intelligence given constraints like size and energy.
We don't really have the tools to reason about this stuff yet. Exciting times.
But, it could happen for a coding-focused model, or an accounting-focused model, etc. most tasks only need a subset of the total model to be done effectively.
Could you not say the exact same of image gen models? For those that haven't kept up with that domain, you can now efficiently run high quality image gen models on any plain old video card, with phenomenal results.
Weather conditions, ground/runway conditions, obstructions, mechanical issues. There are tons of edge cases with huge consequences for error. Same reason self driving is not a solved problem, but with dramatically less tolerance for error.
But it feels like planes have a huge advantage in that they all have trackers constantly broadcasting their location to one another. Almost like they are working as a team compared to FSD systems which are limited to a small range of self collected info from vision/lidar.
Not that it can ever be fully automated but it seems like much easier problem to solve.
Doesn’t meaningfully change the answer. There are still tons of edge cases with catastrophic failure modes. The transponder won’t tell you about the alligator on the runway, a vehicle where it shouldn’t be, unexpected braking action, a mechanical problem, or a plane not doing what it was cleared to do. Tracking the planes is the easy part.
The reasons we haven't automated air traffic control are similar to the reasons why we still have pilots in the cockpit even though we technically don't need to some of the time.
The thing is... not everything you need to track, not even all airplanes, will have transponders.
Sometimes the transponders will be faulty, sometimes you have to direct a broken aircraft safely to the ground with no transponder available.
And sometimes you are doing ATC for a zone that does not, in fact, require transponders (those are required only for the strictest controlled zones, not in general)
This is a slippery slope, where the number of operators can be allowed to gradually decline until eventually it’s just AI. And given the unreliability of LLMs, this is a grave concern that will undoubtedly lead to the enshittification of critical services.
160 degrees for a chicken breast is perfectly safe. Considerably lower temperatures can also produce microbiologically safe chicken when combined with sufficient holding time.
165 is a rough rule of thumb, not some magic number like the boiling point of water at 1 bar or something.
At this point I feel like the only factor differentiating SOTA models now is who they’re propagandizing you on behalf of (not considering agentic tooling/state management, etc).
reply