Hard problems are where the money is made, as they say.

A wise man once said: "if you're given an opportunity to cut an amazing deal and you can't tell who's getting screwed, then it's probably you"

0: https://pestakeholder.org/news/trump-admin-bails-out-private...

* big banks are trying to get out of their data center loan commitments, even selling that debt at a discount. From the article:

> According to the Financial Times, major lenders are already scrambling to offload pieces of massive data center loans through private transactions, risk transfers and synthetic structures. The reason is simple. AI infrastructure borrowing is reaching sizes that are beginning to choke the arteries of the financial system itself.

* there are real questions about long-term liquidity and capital capacity across the entire VC ecosystem. Ed Zitron estimates that the available capital for all technology VC funds will be fully exhausted within roughly two years if current spending levels hold steady. More money has been spent on AI in the last decade than the Manhattan Project, the Apollo Space Program and the US highway system combined[1]

* short-term success of these new data centers coming online is heavily reliant on steady fuel prices since hooking up to the grid can take years and many burn diesel generators while waiting for grid access. If the war in Iran drags on, high fuel prices will continue to ratchet up the cost of data center operations.

* public sentiment around the economy was largely positive heading into the collapse, whereas we've been in fairly consistent state of economic uncertainty for years now. Affordability was not a topic of conversation back then and a majority of Americans are unhappy with the direction of the economy in 2026.

0: https://www.investing.com/analysis/the-ai-boom-is-starting-t...

1: https://www.aljazeera.com/news/2026/2/19/visualising-ai-spen...

That's what scares me the most about this stuff. People see the generated output and get wowed by it, then notice the issues and think "they'll get that problem ironed out", when the truth is that we're gonna probably be dealing with some form of hallucination forever. I'm not convinced at all that we're gonna get past this problem, we're just gonna get more comfortable with letting other people find the bad output.

Sure, that is until each government's dataset is interesting enough to the other to facilitate a data-sharing agreement.

There's gotta be an internet "law" that says something like "Eventually, the data you volunteer to a benign 3rd party eventually winds up being used against you by someone". This is short-term thinking at it's finest.

Have you ever driven on a cobblestone street? There are a few in the city where I grew up and it's pretty obvious why we don't build that way anymore. It's like driving on an uneven dirt road, you're lucky to get above 25MPH consistently lest you want to risk damaging your car.

I've shared before about hiring a local plumber for a remodelling job who replaced their office assistant with a voice AI that could not understand my address (which was not complicated). I rage-quit that call and found another plumber by the end of the week.

It took the first plumber 3 weeks to check his AI call log to find out that he lost a 5 figure job, reaching out to apologize and see what he could do to earn my trust. In that time, I got another quote for substantially less and was already getting the work done.

That's my experience with small business AI solutions, truly terrible experience and I had zero sympathy for the first guy. If I'm paying you all that money to come mess with my plumbing, I want a human on the other end I can talk if something goes sideways.

I think execs are vastly underestimating the damage an apathetic engineering org armed with AI can do to their platform. The short-termism can (and I think will) come back around when they foster an angry culture with a huge token budget.

It reminds me of Woodstock '99, when in order to keep an angry, hungry and drunk crowd under control, the organizers planned a candlelight vigil for Columbine and gave the crowd real candles. That went over about as well as one might guess.

The problem is the scale of investment. For example, 3 companies own 19,000 houses, or 11% of the rental market, in the Atlanta, GA area[0]. The market is being "manipulated" in that investors with vast amounts of cash are outbidding folks trying to buy a primary residence, and by owning tens of thousands of homes, you exert a measure of control over the housing supply (i.e. "manipulation").

I don't believe that companies should have the ability to own so many houses, and the easiest way to curb this is to progressively tax investment property to the point that isn't financially feasible to own tens, let alone thousands, of homes.

0: https://news.gsu.edu/2024/02/26/researchers-find-three-compa...

Comments URL: https://news.ycombinator.com/item?id=48099172

Points: 3

# Comments: 0

The modern javascript ecosystem is a perfect example of what happens when everyone tries to rebuild from scratch and it's a nightmare.

Until they file an S1 to go public and show the world the books, take everything they say with a grain of salt. The amount of financial engineering going on in this space is astounding, and I'll believe it when I see an objective 3rd party release an audit confirming this claim.