borissk

If this is true, than Anthropic, Google and maybe OpenAI models will keep getting better and better and everyone else will be left in the dust - as they won't have access to so much customer data.

The employees are hoping to become very very rich after the IPO and after they are allowed to sell the shares given to them - risking a likely multi-million dollar pay back to leak a model that will be superseded by publicly available models in a couple of years is not a likely decision.

We are really getting close to singularity - the pace of LLM improvement is constantly accelerating.

If China switches to war time economy, they can produce very easily 1M Shahed like drones every month. Plus many millions of FPV/AI controlled drones. Plus massive number of missiles, including hyper sonic and supersonic. They can kill everyone in Taiwan with just drones and missiles if they decide to commit to it.

But given the advances that China has had in EVs, drones, solar, batteries, wind turbines, AI, nuclear energy, smartphones and other advanced industries IMHO it doesn't make sense for them to start a war right now. Better to keep growing their industry and exports and take over Taiwan sometime later.

What do you think about Cursor?

Interesting. When reading dystopian SciFi books, about a future where big corporations are above state governments and dominate the world I couldn't quite believe it. But such future becomes a lot more believable now.

The big tech is going to be one of the big winners from Internet Access Control. This will give them a more reliable way to link a user account to an actual human being - a link that can be monetized in a variety of ways. All kind of political regimes can use such regulations to enhance their control of the population. And the loosers are going to be the Internet users and small companies.

The unfortunate true is IAC is coming to most countries in the world, no matter how much the Hacker News audience hates it...

With $5k you have to make compromises. Which compromises you are willing to make depends on what you want to do - and so there will be different optimal setup.

Can even network 4 of these together, using a pretty cheap InfiniBand switch. There is a YouTube video of a guy building and benchmarking such setup.

For 5K one can get a desktop PC with RTX 5090, that has 3x more compute, but 4x less VRAM - so depending on the workload may be a better option.

The governments may have by that time armies of drones and robots, controlled by a few loyal people or AI.

Those people may have armed robot guards.

The counter argument to that is what happened with horses. Since domestication every advance of human civilization lead to having more horses. Until cars were invented and improved - which eliminated overnight 90% of the number of horses used.

So the fact that in the past new technologies have created new jobs is not a guarantee that AI will create new jobs.

On top of that have a look what happened say during the industrial revolution in Britain. You'd have a village with 2000 workers producing clothes or materials to make clothes. A rich guy opens a factory in that village that employees 200 people and produces more than the whole village before that. 90% are unemployed, the 10% that work in factory have far worse working conditions. Studying graves from that time shows the height of people went down during the industrial revolution - as the conditions in the factories were far worse than what they had before. Hence the luddite movement - but as the reach people owned the mass media, the luddites were portrayed as crazy. Eventually, many years later, new better jobs did appear.

From the decoherence / Many-Worlds view: No collapse occurred. Only entanglement happened.