Johan Strömqvist

Cellular

[Johan Strömqvist @strom_qvist]

@sytelus @anshulkundaje AI will get better. But biology is on another level.

We need new datasets with context across scales, molecules to cells to organisms, and perturbations where we change genes and observe what actually happens.

Wet labs aren’t going away.

[Johan Strömqvist @strom_qvist]

@Math_files ”Fun” footnote: when Kovalevskaya became a professor in Stockholm, August Strindberg called the idea of a female mathematics professor a “monstrosity.

[Johan Strömqvist @strom_qvist]

@davidpattersonx Why not judges?

[Johan Strömqvist @strom_qvist]

@JerrysOrd @LinnSamm @BuschEbba Jakob, Jesus bror enligt Paulus, stenad på initiativ av översteprästen Ananos enligt historikern Josefus. Petrus tros ha avrättats i Rom under Nero enligt tidig kristen tradition (1 Clemens, som dock inte finns i bibeln).

[Johan Strömqvist @strom_qvist]

@juliet_turner6 Why not bacteria then?

[Johan Strömqvist @strom_qvist]

@PeterDiamandis Eventually, sequencing becomes a sensor, not a test, continuously reading biology through DNA, RNA, proteins and beyond.

[Johan Strömqvist @strom_qvist]

@stijnnoorman Complete freedom sounds mildly illegal…

[Johan Strömqvist @strom_qvist]

@pmddomingos If by “done” you mean formally verified, then… sure?

[Johan Strömqvist @strom_qvist]

@VS23220279 @wfoxiii @0xLuciferCore

[Johan Strömqvist @strom_qvist]

@bryan_johnson @bryan_johnson Have you experimented with extremely deep conscious relaxation to reduce sleep need?

If longevity is measured in healthy conscious hours, not just years alive, safely needing less sleep could effectively add years to your life.

Seems testable.

[Johan Strömqvist @strom_qvist]

@wfoxiii @0xLuciferCore 11⁵¹¹⁸⁸ beats it…

[Johan Strömqvist @strom_qvist]

@IntuitMachine @realBigBrainAI Maybe it poses a harder question: if both humans and AI can step outside a formal system, what tells either of us which new axioms to trust?

[Johan Strömqvist @strom_qvist]

@AlexCarten_ This raises an even deeper AI question. The Axiom of Choice is independent of ZF, it can neither be proved nor disproved from those axioms. So what should an AI do when logic itself cannot decide which assumption to make? At some point, mathematics is more than just deduction.

[Johan Strömqvist @strom_qvist]

@juliakw__ Interesting connection to the Axiom of Choice: it is still an open problem whether the Nielsen–Schreier theorem (“every subgroup of a free group is free”) is equivalent to AC over ZF. We know AC ->NS, but whether NS ->AC remains open. Maybe one for AI to crack?

[Johan Strömqvist @strom_qvist]

@FredrikHjelm4 @pjparson A bit more complex.
If VCs are pitching AI exposure to their LPs, they will naturally prioritize AI-centric companies.

That may be the right allocation of risk capital, or not, but it will inevitably shape which types of companies become successful if VC backing is required.

[Johan Strömqvist @strom_qvist]

@HinataMotivates Simulated it: EV ≈ −$0.2/game in the adversarial setup.

Ballmer (or rather von Neumann) was right.

[Johan Strömqvist @strom_qvist]

@HinataMotivates You lose more often than you win… and still come out ahead.

That game is +EV (~$0.20/play) if you play optimally.

[Johan Strömqvist @strom_qvist]

@karibuwu @HinataMotivates I initially thought I could get around this by varying my strategy.

But you’re right, in a minimax sense, that doesn’t help against a sufficiently smart adversary.

[Johan Strömqvist @strom_qvist]

@karibuwu @HinataMotivates Simulated 100 games -> $29 total.
EV is positive. Scale that up and it only gets clearer.