833 karma · joined November 18, 2025
what do people think the billions of billions of pattern matching used in ads will be used for?
people think 'anonymous' credentialing will work here?
they've captured scroll patterns, typing patterns, language patterns, all sorts of fingerprinting.
the game unfortunately is basically already over.
LLM's literally wouldn't work without the sum total of knowledge (in the forms of books and other copyrighted content) being used as 'training data' for these LLMs.
The 'bleeding edge' LLMs required many things, but: 1 Tech innovation ('attention') 2 Lots of compute 3 Data 4 Pre + post training
#4 doesn't happen without #3.
It's pretty obvious at this point that the major providers have stolen vast amounts of #3 - they have paid nearly 0 of the creators.
We can argue about the impact (I'd lean net good) vs. the cost. But arguing there isn't a cost is a bit silly.
Do they have MacBooks in the US that run the queries and stream the outputs back to China?
Who do you think the buyer of retail sentiment and ChatGPT queries are?
My understanding was the main users of your bots will focus on U.S. equity and futures markets. Either way - obtaining the kind and quality of data to find edge (even the Indian Markets one) is likely beyond the infrastructure of most people. do you agree? selling options in overpriced names, for example, requires careful risk and sizing calibration or you could get wiped out before your edge shows up. Or the regime could change, etc. etc.
Either way my guess is most people using this service will be looking to trade things they can like US equities (how many HN people can legit trade Indian Options?).
Presumably the 'small' edges to the degree they existed in the first place have been worn away.
You're paid to hold risk, innit?
Probably taking hate from both sides - OpenAI / Claude fans who are undercutting its moat. Chinese open-model fans that want it to be cheaper.
But it's a genuine accomplishment to hit those benchmarks and offer a reasonable plan?
Bizarre reaction TBH.
89% of these projects are - in fact - doing well. But that number is decreasing. The net result is less supply of public housing in one of the richest states in the entire world.
I'm not sure what your proposal is?
There has been massive public investment and popular support to cause a revival of sorts in the city and is a success story.
Go look at some photos from like 2010-2014.
You can plunk $10-$20K and get land and a homestead in dozens of states.
Agree with the other comment that overbuilding is a reasonable strategy, but if you look at Detroit downtown (mid 2010s) having an overbuilt downtown is bad too.
It's a hard problem.
But their paid plans I'm not sure yet - planning to subscribe and can let you know.
Almost no chance it will be as generous as OpenAI though. They just don't have the money :-)
* ~37K affordable housing units (baseline) across ~400 projects * 89% rent collection rate (down from 90.6% in 2023) * That's 600 units that went 'delinquent' in 2024 - assuming a $24K 'base' rent (just a guess) that's $15M in lost rent. * Deeply troubled projects (that can't survive without this rent) are at 11% - seems like the inverse. * Cumulative arrears (unpaid rent) of $500M
Here's the problem: * If no one had to pay, no one would. * We've tried free housing before - it suffered tragedy of the commons. Not paying means no ownership means subjugation to the worst actions of the worst members of society. * The projects fall into disrepair, there's no way to bring them back, because they won't be maintained.
Landlords aren't a great solution to the problem to be sure. They can be greedy and heartless.
The bigger problem is the bid up of asset prices - aka private equity and class warfare. As soon as you switch (from renting to owning) your incentives immediately shift.
There doesn't actually seem to be a way around this. Taxing to spend on rent ironically makes the problem worse because you just transfer the money into the cash flows of the owners.
Anyone thinking there's a simple solution to this problem hasn't spent enough time with the problem.
This gets you that in a nice neat package, without the underlying tinkering mechanics.
If (big iff) the usage mechanics work out, then this is actually a really good anti-big-model strategy.
They'll be incentivized for your success, not token-maximizing for their investors.
The team is super smart too. What's not to like?
Wishing them the best on launch.
Just a question of whether it delivers the goods.
Inside tech, AI Slop has WAY, WAY, WAY up. Slop replying to slop. Deepslop.
Deepseek and GLM do slop just as well as Opus 4.8 and GPT 5.5 at a FRACTION of the cost.
Frontier AI still cannot do basic valuable things inside companies: * Write non-slop coherent sentences * Come up with 'novel' plans for features * See whether a design is 'good' * Define architecture from a systems perspective with constraints * Tell you what's important
They can absolutely execute mid-level code changes and reviews.
The reality is tech people are trying to use AI to do BOTH and it's costing them big right now.
AI is best seen as bureaucracy, at a personal and company scale. It's become EXPENSIVE bureaucracy (https://attention.substack.com/p/the-age-of-ungovernable-ai-...)
There are some companies whose work depends on efficiently optimized bureaucracy: financial markets will continue to pour millions if not billions into tokens, because the incremental intelligence on the datapoints is really worth it.
For 'glue' between previously hard to connect disciplines - this is also massive improvement. Privacy, data integrity, etc. "reviews" will become much better.
But for actually making the product better and organizations better, it's not clear that tokenmaxxing is an appropriate or even worthwhile strategy.
From experience it instead seems to amplify the native differences in ability, such that the bottom 80% waste tons of tokens to generate more and more output to seem like the top 20% of performers. It can be exhausting to separate the increased amount of wheat from the slop-generated chafe.
This is a losing proposition for the AI providers, and the companies. And will start to really punish the incremental high-priced tokens outside of a very small wedge of industries.
You really gotta hope the next 'versions' of models produce, or the economics are very, very cooked.
we literally have Face ID and passkeys everywhere.
they'll get it right. there are secondary verifications.
the problem is they lose out the young generation, and the cool.
existential risk. big lobbying effort like this.
it's not a dilemma.
Congrats!
If R&D costs don't go up - where does the moat come from? Cheaper players catch up with 'good enough' and will erode their revenue. Most of human tasks just don't require that much intelligence.
They're racing toward 'superintelligence' that recursively self-improves.
No indication we're anywhere close to reaching it.
Going to be an interesting year to say the least.
Some thoughts: 1. US Spending on R&D has gone up from $50B -> $1T annually, and from $3B -> $115B on purchasing power terms 2. The labs rely on government grants, which are hard to get and typically awarded 'equally' or 'by gatekeepers' 3. There is and have been massive scandals that question academic integrity - reproducibility, fake data. The scientific community has done almost nothing to change its mechanisms. 4. It's not clear to me what we've 'got' societally from these studies as a whole 5. The administrative burden to even do science has gotten too far out of hand 6. You can't 'fire' researchers
Research and science is a fundamentally 'good' thing; we should encourage it.
We may need to shake up the way that it's done. Yes science + R&D is long term focused, but it doesn't mean it can't be reformed.
That the article centers on hand-wringing over 'my government grant is gone!' instead of 'you're cutting this critical research that will save lives' without any discussion of 'science' needing to reform unfortunately highlights the core of the problem.
1. Too much inequality, and the risk is bloodshed. This is most of human society. 2. Too little, and you get communism, repression, or European stagnation. 3. U.S. is in the goldilocks zone. The goal is to make it so most people in society don't give up (Japan had this happen).
USD is a bit of a red herring. It does make our capital markets the most valuable in the world, and it does generate nearly endless demand for the dollar, but that's actually based on the wide scale belief that the U.S. is far and away the best arbiter of the world economy. The EU had a shot at that crown and absolutely wrecked it via over regulating.
Travis Kalanick did far worse with Uber. He illegally spied on political opponents to Uber's expansion. As far as I know he's still out there free as a bird.
Private equity is a different problem, which hits closer to the real problems at the edge of the market where prices don't signal reliable information.
We've known for over a century monopolies are bad, but the actual finding is more like concentrated pricing power for goods with inelastic demand is pretty bad. PE exploits these corners with devastating efficiency and often bad results.
You can't predict these systems at all!
If that's your premise, and you suggest that an authoritarian country winning on AI recursive self-improvement does not have the possibility to completely reshape the world into an authoritarian regime... I'll suggest you're falling prey to the exact 1st sentence of the article and making a political point, not a logical one :-)