Modern media is on this spectrum (infinite scroll, algorithmically tailored, short form video).

When people can get dopamine by pressing a button it's easy for them to trade a rich and diverse set of values for that button. When people can get dopamine by pressing a button, it's hard to maintain a rich and diverse set of values.

Values are genuine interest (want) in taking actions and patterns of actions.

Its more likely that they simply value all the convenience they listed over having their own backyard, larger living space, living in a quiet safe neighborhood for kids to run around in, and being surrounded by nature.

I mean that if you accept that the market demand for software development is evolving and you evolve with it you won't end up starving on the streets.

If you assume it's inevitable that you will become useless and starve then it's a lot more likely to happen.

There is a concept in racing "don't stare at the wall". It means when taking a corner at top speeds of you focus on what you're afraid you will crash into you're more likely to crash into it. So instead the idea is to look at where you want to go (around the corner).

So what I'm saying is that if you accept the universe for what it is and focus on what kind of life you want to have and what you need to do to have it you're more likely to get it.

It seems like there is unlimited demand for software, and software engineers who maintain skills relevant to the current market will find work.

For all of history the environment has put evolutionary pressure on living things. If you choose to make a deal with the universe that you get what you want without having to adapt to changing circumstances then you're going to get a shitty deal. If you embrace what's outside your control and adapt gracefully you will get a good deal.

What that means in this particular case with AI and your skills as an engineer I can't say. I don't mean anything specific by it. My point is more so that a big source of unnecessary suffering is wanting the universe to be something that it's not.

So far it seems that we can't all be good/capable at everything.

Maybe you write all your own code and cook your own meals, but do you do your own taxes? would you defend yourself in court? would you fix your car yourself?

Some people may answer yes to all the above, but most people outsource most things. Many people can't afford to outsource to other people, but can afford to outsource to LLMs.

Again it comes down to the job to be done:

- does the person want to develop or practice a skill, or do they just want it done? What is their budget, and how important is it to them to minimize certain risks?

If someone spent a lifetime mastering woodworking with hand tools, and then was shown a couple very early rudimentary power tools (lacking safety features, crude features, etc) they would rightly conclude they weren't useful. The artisan can do better work faster with less risk of dismemberment without them.

Prior to LLMs, the world's demand for good software was bottlenecked by access to competent software engineers. The people who want software just want it, they don't care about the craft. They have a different job to be done than the engineer.

An example this reminds me of is a jobs to be done theory thought exercise:

Two different first time home owners need to store yard working tools in their backyard, and determine they need a shed. The first one cares most about minimizing the time it takes to get the shed. The second one has some special constraints to deal with AND also wants to start developing their amateur construction skills. They both need sheds, but they have different values, so:

- the first one buys a shed-kit made of plastic panels that can easily be assembled in 20 minutes.

- the second one buys a power saw, power drill, tool belt, saw horses, lumber, screws, metal roofing, etc and builds a custom shed from relative scratch over a few weekends.

Another example is getting take out vs cooking the meal yourself. There are many many examples.

LLMs are already useful to many. They are also not useful to many others. To assume they aren't useful to anyone just because they aren't useful to you is a sign of absent cognitive empathy. Not acknowledging that other people have other priorities and values they are equally valid to your own.

I can understand how they would be against competitors distilling their models and for having access to competitor open weight models.

At first glance it seems like a pragmatic answer: "banning open weight model use by US companies doesn't help", but to what question? What if those open weight models are poisoned and primed to create backdoors in US companies? If the government has intelligence this is a credible risk, what should they do about it?

Overall it feels like the letter is conflating a bunch of motives, some of which may be opaque, and at face value doesn't seem logically consistent.

That said, I would expect SpaceX to grow at a compound rate much faster and for longer than any legacy competitor such as Lockheed.

The point I was making was based on a reference to their market cap when it was $20bn in 2018. At the time there was overwhelming and biased negative sentiment about how it was over-valued at the time that wasn't based on a sound analysis.

Funny enough, I think the answer to your question may be in what happened back then.

In 2018 Tesla had a dominant technological advantage in EVs and a strong brand. There was overwhelming anti-tesla investment propaganda, and history's largest short position against them. When a company is over-shorted and then increases in intrinsic value instead of going bankrupt, those shorts covering will put further upwards pressure on the share price. So for at least a while, maybe a long while, the company will be over valued to correct for the previous over-shorting.

It was the same thing with Tesla. Everyone screaming when Tesla had. $20bn market cap how they were extremely over valued because their car sales revenue paled in comparison to the other car manufacturers and yet their market cap was in the same ballpark. It was such a joke and so obviously unprofessional / dishonest. The traditional automakers were all nearly insolvent (tens to hundreds of billions of debt) and had stagnant growth. At the time Tesla had no debt and was growing consistently ~70% a year. Anyone willing to actually analyze and see the truth and not just read headlines and repeat them could have seen it.

The first is a subtle implication that the share price of SpaceX is dependent on the short term feasibility of data centers in space. I don't think it is. Its more dependent on public sentiment and hype, which is detached from the truth and can have its focus directed elsewhere as the company engaged in many activities. I would also argue that the company's culture matters more than any engineering specifics, especially with how diverse their assets, competencies and offerings are. Maybe you didn't mean to infer this though.

Second, if intelligence is something like electricity, in that it's fungible and translates fairly universally into value, then it's safe to assume civilization will pursue expanding it endlessly (and will never have that demand satiated) like a force of nature. If this is true, would we rather have 1000x or 1,000,000x the data centers we have today within earths atmosphere or out in space somewhere? Personally, at that scale I would prefer them far away. (Not even in LEO)

There is no core physics reason why AI data centers in space can't work. It's just really hard engineering. Kind of like what reusable rockets looked like 10 years ago... All the experts were nay-sayers.

Short term, who knows, SpaceX could struggle financially and be a terrible investment today. I have no idea. It seems overpriced to me now and it did at IPO. But on a 10 or 20 year investment horizon, it looks a lot more interesting. I don't own shares but maybe dollar cost averaging in at some point isn't a bad idea especially if the price comes down further and as part of a diversified portfolio.

Maybe I'm being naive, but as someone who spent 2015-2025 fully bought into the idea AI would replace all human labor within ~20 years, my updated take that it won't has a different flavor than coping.

A bit contributor to my perspective was studying the physics of the brain, and a deeper understanding of the very hard economic constraints in the current LLM supply chain.

Re the brain, we compute information in a fundamental different way that is over 1,000,000 more energy efficient than an LLM. We are also around 1,000,000 times more sample efficient at learning.

There is a lot of details in terms of specifics between complexity in human brain architecture (variety of synapses, diversity of neuron types, etc) that have no material equivalent yet.

And on the other side, understanding how hard it has been to build out euv lithography infra, and to completely retool manufacturing for a new paradigm of computing will require so much coordination, a super AI could have all the answers but we also have to deal with getting a bunch of stubborn people to align.

Some people are just coping yes, but if you really zoom into the details on both sides of the equations the gap between where we are and what it would take to dramatically surpass human intelligence is vast.

At the same time I think people overestimate how hard some kinds of knowledge work is (in absolute terms, since it's new and we haven't evolved into it as much), and how economically impactful human augmentation with weaker narrower AI can still be. (Lots of automation and prosperity but without such extreme disruption)

I think in the very long term all bets are off, but we will have plenty of time to evolve our economic systems before we get there.

I agree with your sentiment (about the noise), however I think this over simplifies it a bit. We may get AI that is super-human at frontier research and dramatically accelerates the pace, and still have to wait decades before it disrupts the job market (or maybe never displaces all work).

For one, the answer may depend on material science and chip manufacturing that can take a very long to build out a supply chain for even with super AI help.

And we may just find that the human mind is way more capable than we thought and even with accelerating research it's just a harder problem than anyone expected, even algorithmically.

I expect it to be a bit of both, and from ~2015 - 2025 I was in the "AI is coming for all our jobs" camp. My perspective changed last year after doing a deep dive into latest science on the human brain. (I've kept a very close eye on AI dev progress for 12+ years.

Here is the point from the comment you responded to, copy and pasted for your convenience. Please prove to me that spoon feeding you makes a difference:

> You could make an argument that agriculture is different because we need agricultural products to live. But we don't need those specific products to live (alfalfa, almonds, etc), and they could be grown far more efficiently if water were priced by market rates.

Not maybe, definitely.

> it possible to get a specific teacher's feedback? Not really.

Yes actually it definitely is.

> Is it possible to guarantee it will be productive feedback? No, especially if the student has to/gets to interact with it.

Yes it is possible to guarantee it will be productive, and far more consistently productive than what teachers can achieve.

There are going to be educational contexts where LLMs can't provide productive feedback (because LLMs aren't relevant to the learning objectives). There are also many contexts where they are exceptional at producing productive feedback. Especially in grade school and qualitative undergrad courses.

I am pretty sure what's happening here is you are conflating LLMs with ChatGPT and other chat interfaces. That's a bit like conflating an internal combustion engine with a tractor, and then basing your experience with tractors on an opinion that busses can't exist. Indicated by this thing you said: "No, especially if the student has to/gets to interact with it.". It seems like you haven't considered that LLMs can have any kind of guard rails or custom instructions applied to them, can be packaged or constrained in how the user interacts with them.

An interface can allow a student to submit a draft, get static personalized pedagogically-optimized feedback tailored to the teachers criteria, learning objectives and reference material, without any way for the student to get any other output.

I find it both funny and a little irritating how confident you are that this isn't possible, because I've seen it with my own eyes used in both graduate and undergrad contexts to great success.

In a way you have answered my original query, which I am grateful for: "Why aren't all educators tripping over themselves to use LLMs to maximize access to timely constructive feedback?"

You're indicating the answer is because most educators are confidently incorrect about LLM capabilities. Plausible I guess.

My gut still says it's easily misinterpreted and can be used as justification to do something that shouldn't be. It has something to do with scope of applicability, but to be fair that's true of my other "good" examples as well (e.g. water boiling at different temperatures / pressures).

My instinctive "hold on a minute" was the assumption "everything in moderation" was not a good generalization when applied to everyone as individuals rather than more broadly/universally.

I suspect others that took issue had a similar misalignment in scope. (Fentanyl in moderation for a drug addict is not sustainable, fentanyl in moderation for a non-addict recovering from surgery, probably good)

I feel like I've stumbled into some philosophy tar pit - questioning the meaning of generalization, moderation, scope, good judgement and "everything"... so I'll just concede for now, despite feeling uneasy about it.

The real crux is not grounded in chat bots default behavior but the technology's capability: "Can LLMs provide access to timely constructive feedback in specific educational contexts?". The answer to this question is definitely yes. If you think the answer is no, then my guess is you haven't made an honest effort to try, or you just want the answer to be no and aren't interested in the truth.

2. Designs require advanced simulations which benefit from AI both in building the simulations and running them.

3. Material science benefits from models learning how particles interact and materials behave in certain conditions so we can predict which materials are worth producing.

These are odd the top of my head and I'm not a fusion physicist. There's probably a lot more examples.