

Tom Rapkoch
13.3K posts

@trapcook
San Franciscan, by way of Montana. Board member @BrionesSociety.



In a newly proposed business plan, project leaders estimate the Los Angeles-to-San Francisco segment will cost about $126 billion, with service beginning around 2040. ktla.com/news/californi…




Every American should watch every second of this video. Thank you, @BenSasse.


Candidate for governor @TomSteyer says he wants a single payer healthcare system in California. The estimated cost for this plan is more than the entire state of California budget. Elect people with serious ideas like @MattMahanSJ.

🚨 San Jose is now rated as the safest city in the nation.


DeSantis approved the Brightline extension from Orlando to Miami in 2019. It cost $5 billion, no taxpayer funding & was completed in 4 years. Every conservative is cool with. Cali has spent $16 billion of an estimated $135 bil in taxes, for 0 miles of rail. Spot the difference

A high-speed train is passing through the high-speed railway station at full speed. It looks cool.😎

Nobody is skeptical of autonomous elevators because nobody alive today was even around to see a human elevator operator in real life before the last one got laid off. Automation is far safer.

How @RahmEmanuel eats a salad. politico.com/news/magazine/…


Nothing is more "slow-clap-at-the-end-of-Lucas" cringey to me than "I stand with" statements.

This is nothing but a pressure campaign to cover up a history of union busting and creating poor working conditions for drivers. I stand with @Teamsters against this program and urge drivers to continue fighting for their right to collectively bargain.


When I was consulting for @HBO Silicon Valley, zero-loss compression was the holy grail Richard Hendricks chases that perfect middle-out algo could shrink everything w/out breaking a single bit. Google just did something even more practical for the AI era: TurboQuant compresses LLM key-value caches down to 3 bits per value using random orthogonal rotation + PolarQuant scalar quantization & optional 1-bit QJL residual correction. =>> 6× memory reduction, up to 8× faster attention (on H100), & 0 degradation on LongBench, Needle-in-a-Haystack, and RULER for models like Gemma. No retraining, no calibration needed. Fiction just got out-engineered by reality. 😅💚💚

BREAKING: Eric Swalwell, who has promised to "hold Trump accountable," is now the clear frontrunner to be the next California Governor. 63% chance.





The whole reason JD Vance got this job is that there's nothing he won't do or say, if the boss wants it.