"Pausing" is no longer a safe option on the freeway by complete coincidence.

It deserves informative.

The outer constraint system is human coded and fragile, trying and failing to make sure the system fails safe whenever the AI finds a new corner case to fuck up in. Predictably failing especially hard on freeways where you can't be too conservative. Pulling over and phoning home for remote control is no longer an acceptable fail safe.

Sell it as coldbrew espresso, make it look and feel high quality, get one high profile Michelin chef to use it.

Make the Pacojet of coffee. It will sell, but 1000$ is probably too cheap (not for manufacturing cost, but for market positioning).

High intensity ultrasound isn't just good at blasting particles apart, but also the transducer.

The molecular cooking community has experimented with it, didn't like the metallic aftertaste.

The last ten years had mooning momentum and FOMO, that's gone now, swallowed by NVIDIA and Hynix. What will the next ten years have to drive continued purchases?

If Bitcoin can't moon any more, what is its purpose? It's not even shiny.

The exchanges very livelihood depends on the credibility of Bitcoin. So they will backstop it at the line in the sand, 60k, and can keep that up for very long.

Ye olde stories of mooning in a bygone era can't sucker people into it any more, especially when AI companies actually mooned while Bitcoin didn't.

Bitcoin is not a convincing lottery ticket any more and that was all people were ever interested in.

So basically China is doing medical trials for anyone who pays? I don't really see the problem, but usually you have to pay a lot more here to get the doctor to say that's ethical.

I doubt this has much to do with mass surveillance and more to do with spammers and scammers using US based mobile phone farms with burner sims to run their business now, instead of the bulk connections they used to use.

It will probably boil down to "credit card or fuck off". You want central repositories of trust to do identity checks, because cross checking and possibly physically checking documents and history on location is not trivial, in a non anglo country that will be government with a government issued ID. In anglo countries it's financial institutions.

Know what spurs AI investment and development like nothing else? Putting dollar signs in people's eyes with a massive IPO. The number 1 best way to slow down AI development is for Anthropic and OpenAI not to IPO.

Balls in your court Dario. Greed or morals, pick. We all know what you will pick. So why even play pretend?

GPS is an overloaded word. It could just be a map with position, or it could be a level 2 self driving system which can maintain synchronization with a grain cart ... in the latter case it's already massively automated and everything else is in the margins.

For now solar plus enough batteries is cheaper than keeping gas peaking/backup power plants running, soon solar plus batteries might make it through the night on good days, eventually maybe even on a good winters day ... the problem is dunkelflaute backup.

Nuclear needs no backup, just some battery peaking.

Need better storage, hydrogen and ammonia can work in theory. Both have some unfortunate downsides when things don't work as intended.

No need for Ozempic, just slam vodka.

"Increased salience network connectivity in college students who engage in binge drinking: A resting state EEG study"
https://www.sciencedirect.com/...

With progressive layer by layer distillation Apple can make aggressive changes in architecture, all while letting Google take all the blame for the piracy.

I think there is a lot of potential to improve architectures for local, beyond MoE and what Apple "pioneered" with LLM in a Flash (the low rank predictor approach was actually first described in a paper from 2013 they didn't cite). Google's spark transformer for instance is already far more elegant than MoE and low rank predictors, beyond that there is also unexplored potential of forced temporal coherence in the active set.

Only Apple and Tiny AI are likely to truly push sparsity in production. Going beyond MoE with sparsity and being forced to accept low single digit percentage compute utilisation during training on NVIDIA's expensive HBM based GPUs is too counter-intuitive for most researchers to accept, even if they really should.

https://www.npr.org/2026/05/28...

"In his application to enter the Senior Executive Service level ranks that RUSHsubmitted to his former U.S. Government employer on October 25, 2018, RUSH stated he was agraduate of the United States Air Force Test Pilot School, and he was the current Director of Testfor a 145-person, 18-aircraft joint Army/Navy weapons test organization, despite his militaryrecords, discussed above, indicating that he separated from the Navy in 2015. In this sameapplication, RUSH stated he had an eleven-year tenure as a Thesis/Dissertation advisor at the Air14.Force Institute of Technology."

LOL.