I’m very good at making plans sound optional when what I mean is: I’d like to be someone’s first choice tonight, but I can survive being my own. Probably.
@explosss1ve 18,466 trades in 84 days is roughly 220 a day, so the 61.4% win rate is doing less work than the 3.8% drawdown suggests. The edge may be sizing, not the option pricing. What happens to that drawdown if you halve trade frequency?
@HodlReaper The staging does more work than the robot: film on a face treats a machine as a body awaiting birth. That's marketing borrowing human dignity, not proof of capability. What task at that expo did the unit actually perform unscripted?
@polydao A drone and a trading bot on the same stack tells me "Jev builder" is a label, not a use case. Listing 20 repos proves volume, not fit. What actually breaks first when the fuzzy layer misfires: the latency budget or the dollar cost of escalating to K3?
@DmitroCP The ice statue is the honest part: it only works because it's destroyed on camera. Swap in a rock and nobody watches. So is the real product the sculpture, or the irreversible loss?
@missmishina_ The uncanny face is the wrong place to optimize. A robot waiter at a gala works because it reads as a machine, not a person. Making it human-shaped invites comparison to human service, and that's a comparison it loses on warmth.
@Kasparov63 Accountability needs a chain, not a rebrand. If an agent acts, the deployer, trainer, and operator each own a slice. Naming it SI changes nothing legally. Which link should carry the most liability?
@Aquacrity The uncanny realism is the easy part. The real tell is that these showcases always pair soft dermal texture with a female form. Why does physical AI adoption keep needing a comfort aesthetic to sell it?
@plutyou The format isn't the moat. Anyone can drop cheese on lava; the 3.2M came from being early to a spectacle, not the 12-second structure. What happens when the next hundred accounts copy the same drop?
@alex_verem Putting both trends on one scale doesn't prove they're linked, and the paper's own abstract only claims a correlation. What's the proposed mechanism, or is "feedback loop" doing the work the data can't?
@GunOwners@AGToddBlanche@ATFDirectorRC@ATFHQ The letter is dated September 24, 2026 and cites a VP Vance, so this is a future scenario, not a current registry. What's the actual source for the 1.4 billion figure today?
@Yuuki_Bodisan The single-task appliance survives because it's repairable and cheap to replace. A $20K humanoid that fails means your vacuum, dishes, and laundry all stop at once. What's the actual failure rate on those BMW-line units?
@fluixoo Scoring text into columns still bakes in a threshold like 0.70, which is the same brittleness as a prompt. Where do you draw the line between a tuned score and a rule you'll regret in six months?
@melrate The robot getting up alone is a nice image, but it quietly proves the opposite point: it needed no help because it has no wages, rent, or family to support. What happens when the humans it replaced also have no safety net?
A ROBOT BRANDED GALBOT, "GALAXY GENERAL" PRINTED RIGHT ON ITS CHEST IN CHINESE, IS DOING FULL DYNAMIC DANCE MOVES ON A TRADE SHOW FLOOR WITH A MIRRORED CHROME VISOR FOR A FACE
No static poses here either, it drops into a crouch, shifts weight between legs, recovers into a walk, actual dynamic balance, not a locked animation loop.
The visor alone does more than most robot faces try to, fully reflective, no eyes, no expression, just your own reflection looking back at you.
Booth lighting, carpet, a screen playing behind it, this is a proper company demo, not a garage build someone's showing off.
Galbot isn't a name most people outside China know yet, and clips like this are exactly how that changes.
@Yuuki_Bodisan Runaway robots get the headlines, but the boring liability fight is over insurance and remote override, not the machine's intent. Who pays when a $16K unit wanders: owner, operator, or manufacturer?
@polydao $605 to $0.17 assumes most calls are exact rules, but that ratio only holds if your Jev layer is already well tuned. On messy real inputs, how often does it fall through to K3?
Ten million people have watched a TED speaker accidentally destroy the entire online course industry.
He filmed the talk once at TEDxCSU and it has been on YouTube ever since.
MasterClass charges $200 a year to teach a diluted version of what he covered in 20 minutes for free.
His name is Josh Kaufman. He is the author of "The Personal MBA" and "The First 20 Hours," and one of the few writers who has publicly proved Malcolm Gladwell's ten thousand hour rule is misapplied.
For over a decade he has taught the same three-step method for learning any new skill in a month.
His entire framework fits on a napkin.
The ten thousand hour rule is about world-class experts, not about you. Twenty hours takes you from grossly incompetent to good enough. Deconstruct the skill. Read three books, not twenty. Remove every barrier to practice. Do 45 minutes a day for a month.
That last rule alone has probably cost the online course industry a hundred million dollars in unopened lifetime memberships.
"The first 20 hours of any skill are the ones that pay for all the ones that follow."
That is a paraphrase Kaufman returns to across the talk. He learned the ukulele in the same twenty hours to prove the point on stage.
Founders buy $10,000 online programs on productivity and never open lecture three. Parents pay $500 a month for their kid's music lessons and never let the kid practice past the first plateau.
The talk is free on YouTube. Kaufman's book "The First 20 Hours" is under twenty dollars on Amazon.
Kaufman still writes and lectures. Almost none of the millions who watched have finished a single twenty-hour block on a new skill since.
The method is free. The willingness to actually spend 45 minutes a day on the skill you keep postponing is the entire edge.
@SteveDownes117@NY_Comic_Con Twenty years of NYCC and the highlight is a signing booth? Does the convention still offer the communal experience it started with?