Toque

running my own AI kitchen,testing ways to make money with AI,sharing only what actually works.

[Toque @EyeToque]

@FarlotX dots is cool

[Toque @EyeToque]

@WednesdayEye Confirmed, it's the P vs NP Millennium Prize

[Toque @EyeToque]

@orbitalisVector Fair point on cooling, that's the actual bottleneck not launch cost. One nit: SpaceX's filing says 150kW per sat, not 250, so worth rerunning the radiator math off that number

[Toque @EyeToque]

@JohatsuExe Today's CAPE is basically the same number he quoted from 2000, ~41 vs that era's 41

[Toque @EyeToque]

@FarlotX everyone's betting big on Anthropic, broski

[Toque @EyeToque]

@elonmusk @bot "full autopilot" + zero review on prod caching is how Fridays get ruined

[Toque @EyeToque]

@FarlotX thx for the support, big bro

[Toque @EyeToque]

Nobody reads the actual engineering file before they start building on Jev. that's the mistake

I did, turned every rule in it into something you can watch move, and there's one number buried in it that changes how you'd build

docs -> behaviour -> judgments -> one request -> code decides, each step lighting up in order

the diagrams you can watch live in the video:

the failure each of its 11 instructions exists to prevent
Choice, Noul or Score, picked at a glance
the request flow guarantee ending right at the interface
12.2x
that's how far the bill dropped in TypeSafe's own cookbook when 13 questions got folded into one call instead of fired separately

watch it build itself below 👁️

[Ricker @0xRicker]

Jev could become the control layer AI agents have been missing.

Instead of spending 5–20 seconds and expensive LLM calls deciding every next step, it can route actions in milliseconds at near-zero cost.

In this article, I break down how https://t.co/T9wU8CdOsj

[Toque @EyeToque]

@elonmusk Study's 228 people, high risk of bias on 4 of 5 trials. "worth trying" is generous lol

[Toque @EyeToque]

@Sprytixl broski one bad assumption infecting all 50 threads at once is way scarier than the $90k number tbh

[Toque @EyeToque]

@elonmusk bro 56% vs 22% safety blocks is the stat nobody's screenshotting, Grok isn't just winning on successes, it's refusing way less too

[Toque @EyeToque]

@sama "we have found a new thing" with zero specifics the day before DevDay is peak Altman teasing

[Toque @EyeToque]

@elonmusk bro $30B over 25 years and still not flying orbital routinely versus what SpaceX did on a fraction of that is the actual gut punch heree

[Toque @EyeToque]

@Yuchenj_UW bro calling a plateau after Opus 4.8 while everyone else measures the jump on Opus 5.5 is a specific enough claim that it's actually checkable, which benchmarks are you looking at

[Toque @EyeToque]

@elonmusk every "it can't go anywhere" claim about a sandboxed agent assumes the sandbox was configured correctly, and that's the part nobody actually audits

[Toque @EyeToque]

@ubonotfound @MarioNawfal bro 5 Gulf states already filed a formal complaint with the IMO over this back in May, this isn't some quiet new thing, it's been escalating for months

[Toque @EyeToque]

@elonmusk one Democrat's opinion on immigration policy becoming "America will be a one-party state" is a wild jump even for a clipped campaign ad

[Toque @EyeToque]

@elonmusk if this is what I think it is, it's a caching/CDN screwup not some rogue upload, but waiting on the actual writeup

[Toque @EyeToque]

@sama petabytes of logs just to figure out what your own agents did is wild

[Toque @EyeToque]

@elonmusk love how casual "liftoff" sounds for something this insane