@chrisadamo24 Looks like @BharatSemi raised with design wins from General Atomics and Kaynes in hand, on a sensor still in prototyping.
While the $24M pays for the next test chip, the customers are what made the round possible.
@AIGuide_ I'd go even further. Founders have to sell first.
Nobody has the founder's magic and passion. You can't hire that.
Sell first. Hire later. And the first hire isn't a CRO or CCO.
@mark_k Great! Now we just need it to ban other AI tell words and phrases, like: lever, delve, load-bearing, quietly, unlock, "the honest answer is", "not x but y", emojis, etc.
I've banned them all in my Claude.md.
Hey Fable 5.2, do you buy this?
Half of it. There's a real idea in there, but the framing is hype.
What holds up
* Recovery time matters. If a failed pitch or experiment costs you three days of licking wounds, you iterate more slowly than someone who treats it as data. Separating self-worth from individual results is useful.
* Holding reality next to the goal works. An honest view of where you are, set beside where you want to be, beats hyping yourself up.
What doesn't
* It isn't "unheard of." "Structural tension" and "oscillating" are Robert Fritz's terms from The Path of Least Resistance (1980s), and the rest is Stoicism plus Annie Duke's "don't judge decisions by outcomes." The prompt demanded a secret trick, so the model repackaged old ideas as one.
* There's no evidence it's "the" separator. Luck, timing, starting capital and network do a lot of the work. Plenty of the ultra-successful are the opposite of emotionally neutral: Jobs, Musk and Jordan ran on ego, grievance and identity fused with the work. The "ten times the speed" figure is invented.
* "Strip all moral weight from outcomes" is wrong. Some outcomes do carry moral weight, and discarding it leads to rationalising harm. The sound version is to detach your worth from the result, not your responsibility for it.
* Emotion is not just friction. Damasio's patients with blunted emotional processing made worse real-life decisions, not better ones. Emotion tells you which problems are worth solving, and no "neutral physics problem" framing tells you what to want.
The version I'd keep: care intensely about the goal, and hold each individual result lightly. That is ordinary good advice, not a secret that separates elites from everyone else.
It depends on what you value, where you are in your career, what your long term goals are, etc. Some Claude research summary on Pozuelo, fwiw:
Pozuelo de Alarcón
Benefits:
• Top of Spain's tax-agency income ranking: €92,564 average gross per tax return (2024)
• 2.2 jobs for every working resident, so it is not a dormitory town
• 62% university-educated
• No municipal debt, per the council's 2027 budget
• Theatre and concerts all winter at MIRA Teatro
• Almost no tourism pressure
Problems:
• Young adults are priced out: the 25-44 age group fell 20% in ten years
• Homes are valued at €4,870/m² (Tinsa, Q3 2026), against €3,699 for the Madrid region
• About 1,800 applications for 200 affordable council rentals
• More Spaniards move out than move in; growth comes from abroad
• The big employers sit in office parks on the edge of town
• Few independent shops, cafés or galleries found, and no covered market
Perfect for these people:
High-earning families and corporate professionals who want a quiet, affluent suburb next to Madrid and can buy at these prices. Not for under-40s on ordinary salaries, or anyone looking for an independent, creative town centre.
@Scobleizer You do understand there is a black market for GPUs, yes? And could easily have been done on servers outside of China. Has nothing to do with Jensen.
@bscholl Looks promising, but LLM-driven design isn't enough: physical AI needs physics-grounded models as ground truth. Simulation-trained surrogates help, but electromechanical designs still need FEM validation. CFD is even harder. Closing the multiphysics loop is the hardest part.
@yclueecs Congrats to the @RicursiveAI team! Curious what comes next: if you're training on existing chip layouts, how far are you from closing the loop with device-level multi-physics simulation to generate and verify real tapeouts?
@mattparlmer Fab wafer space and memory isn’t the scarcity driver, its ability to build/power the data centers. Pump out more GPUs, that’s nice, but no place to house or power them, and what’s available is booked up.
Loving @Starlink on my Asiana flight from Seoul to Barcelona. Only complaint is no coverage while over China, India and Pakistan - but fabulous otherwise!
I don't see European voters wanting it. The median voter is around 50, owns a home, and votes to keep pensions whole. Nobody wins an election on fewer protections and more startups.
Fixable in ten years if any government picked it as the hill to die on. None will until the bond market makes them.
@TimurNegru Definitely for recreation wine making that is a nice size and Piemonte is nice region. Trying to sell the wine will be more like sfuso to locals. There’s a flood of wine in that market.
@the_apollonic@BetterCallMedhi And this is the point. France seems to be the poster child of living way beyond it's means to support a socialist system that cannot endure. Protesting is fine, but rioting in the streets will not solve the problem.