rozap

hindsight is the best hedge fund

Even in the first paragraph the author goes into some nuance about what they mean and it's not at all what you're saying? Years ago, I'd work on one or two things at a time. Now I do maybe 10+ things in parallel. I work more hours, "let me just finish up this one thing" I'll say, and I look at the clock and it's later than I thought. At the end of the day, my brain is cooked. Right now I'm motivated by the novelty of this new tool that gives me superhuman productivity, but a few years from now when that novelty has worn off? Well, probably a few years from now we won't have software in any recognizable sense anyway, so maybe it's not worth thinking about.

This entire demo is so incredibly techbrained I had to laugh. These people are really unable to get out of their own heads.

"What can you use it for?" "Uhh....yes, asking investors for more money"

We have done this internally at our company (link in bio). We have an internal simulation game, kind of like the OP. The game was vibecoded, but it can run using the same rules as a city in our actual system. The simulation game is really just a toy that is fun to mess around with, but the rules engine is an actual thing in our product.

My point is that there's nothing wrong with glue, but comparing the pile of shit that is the cybertruck to one of the GOATs and saying "look they both use glue" is a little disingenuous. They are completely different processes, do it in different places for different reasons. The only similarity is that there is glue involved, somewhere.

If you believe that the upside is effectively unlimited because you automate most or all jobs in existence, then it is "rational" but silly valley rationality usually ignores things that are inconvenient. But if you truly believe the upside is unlimited, then rising interest rates won't scare you.

I think the issues are:

1) that timeline has to line up with the broader macroeconomic factors (unlikely imo)

2) anthropic/openai/etc have to actually capture the value (unlikely given deepseek/glm/etc)

3) for all jobs to be automated, there needs to be a huge shift in the social contract and welfare state, which requires competent and functioning government (lol) and taxes on destruction of labor rather than labor itself (lmao) otherwise you get millions of unemployed people and a deeply unstable society, and a lot of blown up datacenters.

I think the value of AI is unlike anything that has come before, I just don't think the big players will necessarily be the ones capturing it. There will be a lot of spectacular implosions over the next couple years, and AI will continue to march along.

Lotus did this on their chassis not because it was cheap, but because it allowed them to use lighter, thinner aluminum which would have been ruined by welding. It was almost certainly not cheaper than welding, they had to add little protrusions that kept all the pieces 0.2mm apart, which is what made for the best bonding. They went to a lot of trouble to make this process work, and it resulted in one of the best sports cars of all time in terms of pure driving experience.

Tesla has very different reasons for doing what they did on the cybertruck. It's very much a form over function. Lotus has always been function over form (to a fault, in terms of their financial position).

I can imagine having kids is super rewarding, but also amplifies the pressure of being a provider probably 100x. In uncertain times, I think a lot of us aren't sure how we'll provide for ourselves in ~5 years, let alone a tiny dude, which plays into the decision to not have kids.

I'm still gainfully employed, and business is ripping, but a huge part of my value as a professional got totally vaporized over the last year. So part of me is extremely glad I don't have kids right now, and part of me is really angry that everything feeling so unstable is leading to our decision to not have them. I know that even in the best time it's a scary decision, but the general vibe is flipping people (like my wife and i) who are at the margin.

Then you can't bill insurers for it. That's where the real money is.

Facebook and its ilk sort were more or less the unravelling of western liberalism. The descent happened before generative ai.

...ok, but he really let it rip. Israel has tried to get every president since Kennedy to attack Iran and nobody was dumb enough or compromised enough to do it until now. It has been bad for a while, and it's naive to think it's not particularly bad now. Every president post 9/11 has been shit, but i can't think of any other ones in that cohort who have had such an acutely negative impact on my day to day until now.

Thurston county WA collects my property tax. Thurston county should not have attacked Iran and caused a surge in energy prices. The county should have hiked rates earlier and more aggressively. Thurston county is out of control!

Government is not a borg. We elected the dumbest motherfucker in the country, he consolidated power, and we got incredibly inflationary (among other things...) policies. Local government is just along for the ride.

Then I don't really understand the argument around restricting everything to a very specific PU formula. Like they dropped the MGU-H to make the unit as a whole cheaper "so that new manufacturers could be drawn in", but then, if there's a cost cap anyway, then what does it matter?

I think you're right, but I do wish the formula was more open ended. The super strict regs were supposed to prevent runaway costs, but now we have the cost cap. So like, why not institute a cost cap for the pu manufacturers and let them go nuts.

I have a chuckle every time the Williams is on the screen with its Claude sponsorship. Why don't they simply ask Claude to make their car fast, are they stupid?

My wife is tired of this joke.

I recently had this discussion with an energetic junior coworker who just learned about TLA+ and thought it would solve all the problems. I think it probably did give the LLM that he was using to do the actual implementation a better starting off point, but the obvious gap remains, which is whether or not the spec (TLA) matches the implementation (elixir) which there's just not a good answer for. I encouraged him to put his TLA code in our docs folder, because as far as I'm concerned, it's just a suggestion, and then write (and understand...) some property tests about the feature.

I think there probably is some value in vibecoding TLA specs and not actually understanding the invariants yourself, but it's way oversold by the talking heads of the tech world, and the gaps need to be filled in some other way if you refuse to write your own code.

It didn't want to tell me in any detail why diesel is $7.50 per gallon

Just because the market reacts to a piece of information to the tune of 6B, doesn't mean the guy is providing 6b of value. It's just a piece of information and the market reacts to what it may or may not mean. Any other goober with an MBA in that role who quits with no notice may induce the same market reaction, even if they were a shit ceo.

Agree with everything you say here.

> I don't think anything here is new or changed

I would argue it's a little new though. I used to write dumb little programs all the time that explored an idea which was probably bad, and in that exploration I often found that there was a better way to do it, or that I didn't know as much as I thought I did, or that another thing already existed that was much better considered than my half baked idea, etc. But there was learning that happened there, so the process was still valuable. Now you can get a working bad idea without learning anything, there is full conservation of ignorance, but a full dopamine hit from "i made this thing". I guess you could argue it's just everything happening at a faster rate, but it feels different to me, and it is pretty eerie.

If AI ruins social media, then that's a huge net good.

It's hard to discuss things seriously when the arguments are comically bad. This whole marketing piece leaves out anything beyond the first order effects, no mention of any sort of class upheaval, no black swan events. All the projections are up and to the right, there isn't even a low confidence scenario where things go poorly. It's hard to discuss an such a deeply unserious thing seriously.

How does that coexist with Capitalism? I just don't see a path.

Regardless of where you fall on the alignment chart, if you are looking backwards from now to the industrial revolution, it's pretty easy to make the claim that from a standpoint of pure technological progress (discounting all other things, like quality of life, environmental impact, etc) that capitalism "won". But going forward, if the need for labor is completely crushed, and capital continues to get concentrated into the hands of the elites, I just don't see how that is viable long term. It seems like for things to shift to the world Keynes (wrongly) predicted, where we all have abundance of leisure time, then there's a massive distribution of the fruits of automation into the hands of the masses, which needs to be controlled by an entity who does not purely exist to acquire more capital, which is sort of the end of capitalism as we know it.

Or maybe AI kills us all before we get there, who knows.

wait you forgot vertex, that one is different

Yea, it does apply. Honestly I'd argue that as you move up the abstraction layer cake, it applies even more. If a team gets a tool which 2x's their output (maybe a higher level language vs writing in C), then the difference in absolute output someone who produces 1 widget per day and 5 widgets per day becomes substantial. If you're given a tool which 10x's your output in a day, then it's even bigger. And of course you can have devs on a team who produce -.5 widgets per day.

Think about the worst colleague you've had, now give them a tool which 10x's their output :)

Silence, we have disruptive direct to consumer peptide startups that need to pump.

The gas truck just came to the gas station on the corner and it's just crazy how big those things are, there's such a surplus of gas right now.

There's a nice extension called unhook which blocks a ton of the YouTube trash, you can configure it to hide different things. At this point, they only thing I see on YouTube is a search box and my subscriptions.

Plenty of people work at Meta.

Yea I figured it was my LSP but when I disabled it, it wasn't resolved. I didn't debug it, I just switched.

It's funny, zed started out really refreshing, and now it's degraded to the point where it's fairly unusable. Regularly it would tie up tons of resources on my fairly beefy laptop. I switched back to sublime and it is a breath of fresh air, I didn't realize how slow zed had gotten. It seems like maybe they lost their way a bit, but I recognize that also they need to make money presumably, and it's hard to do that with a large team and a focused product.