@Tesla@elonmusk Congratulations on 2026.38: refreshed vehicle models and parking visuals for older AMD cars. Intel owners get another opportunity to admire what we’re missing.
Our Intel cars can already display full-screen driving visualizations under FSD. Yet a basic full-screen option outside FSD remains apparently too ambitious for the company that promises autonomous everything.
Nobody is asking an Atom chip to run your fanciest new lighting effects. Give us the existing visualization, full screen. Lower the detail if necessary. This shouldn’t require a technological renaissance.
Is your software team too incompetent to support that, or too arrogant to explain why it won’t?
Are you just waiting for every Intel owner to drop dead so the support backlog clears itself?
“Your car gets better over time.” Apparently, terms and conditions apply to how long you’re allowed to remain a customer.
#Tesla #IntelAtom #202638
I rarely side with you, but on this I’m with you. Open-sourcing frontier models is just giving away years of hard work. Closed-source and powerful models better protect capital and people’s livelihoods. Some Asian countries prefer others to open-source so they can copy it and then call it original innovation.
Just shipped: OddPea! 🎮 Got five minutes? Don’t spend them hunting for a game. Pick a free browser game, chase a better score, or make something just because. 🫛 Try a round: https://t.co/7upRYdfdDp #BrowserGames#IndieGames
@aleabitoreddit He won’t return unless the internal stakeholders are prepared to vote him in with a decisive absolute majority and grant him extra shares so he can actually control the company.
I need to walk back my earlier take that Jensen leaving would mean NVIDIA’s decline. This year he’s been publicly pushing open-source and selling chips to China—quite a shift from his early founding-era stance. On open source: if he truly believes in it, why doesn’t he open-source CUDA? Simple reasons—he doesn’t control the rights, and the early co-founder/CTO’s exit shows CUDA wasn’t really his. Over the last two decades his heavy involvement looks mostly like personal affairs and sales. His departure doesn’t equal NVIDIA’s fall; it could even reinforce its dominance. This isn’t just about possible tech leaks. His open-source advocacy this year would shrink returns for AI model developers who poured real money and their prime years into the work—open-sourcing means sharing a cake others baked. Dressing up Eastern copycat/shanzhai culture as “open source” hurts the tech ecosystem. Otherwise, what’s the point of engineers grinding hard?
Just shipped MotifDraft! 🧵🛠️ Turning a picture into a craft chart shouldn’t mean buying desktop software. Convert your image to a cross stitch, crochet, or bead pattern, edit the grid, and print: https://t.co/xrfYfSx8JK #Crafting#BuildInPublic
Just shipped BaZi Cue! 🚀 Confused by BaZi chart jargon? Get your Four Pillars, Day Master, and Five Elements calculated with transparent rules and zero sign-up. Try it free: https://t.co/9J7hr0XbHW #IndieDev#BuildInPublic
Just shipped LudyPlay 🎮 No more spending your break choosing a download. Play open-answer trivia, a 15-min survival run, or a coastal motorcycle race right in your browser. Free & instant: https://t.co/pvv38Zeavc #IndieGames#BuildInPublic
Jensen Huang essentially is NVIDIA. If the board fires him over something superficial, the company could basically go to zero. I used to work at a headhunting firm where we poached engineers from the most advanced process-technology company at the time. After their successive CEOs jumped to a client company, at least 40% of that firm’s employees followed them. These days, a CEO’s shareholding doesn’t necessarily mean much power—Huang’s personal interests and capabilities are more than enough to make the people who support him change jobs for him. The same dynamic was very clear in the F1 Red Bull team with the Christian Horner situation.
The absurdity of the current media frenzy is remarkable.
We are watching people rush to put Jensen Huang on trial in the court of public opinion before an independent investigation has even established the facts.
NVIDIA is not a one-man startup, but Jensen is undeniably one of the most strategically important executives in the global technology industry. His technical vision, leadership, relationships, and execution have helped build the AI infrastructure powering the modern economy.
A board of directors is supposed to act in the interests of shareholders, not surrender to a social media outrage cycle.
Unless the investigation establishes serious corporate misconduct, a material breach of trust, or actual damage to NVIDIA, replacing Jensen would arguably create far greater risks for shareholders than retaining him.
Let's be realistic: boards don't casually remove the architect of a company's competitive advantage because the internet demands a spectacle.
Jensen Huang's continued tenure may well be the more rational outcome. The investigation should determine the facts — not the outrage machine.
@aleabitoreddit Memory and storage are far more valuable today than in the past because of AI-driven demand. Data centers will keep expanding for years, so walking away from committed orders is a shortsighted move that risks supply shortages and higher costs later.
The recent CapEx surge at the major tech firms already factored in community and livelihood concerns — the spending is inevitable, yet these companies continue to grow, proving AI still has real expansion room, and by embedding local benefits into their outlays they have already begun to better align data centers with residents’ interests while balancing the government’s stake.
What the AI leaders will actually do after the midterm data-center backlash
1/ The midterms just put ~40% of planned U.S. data-center capacity under governors who campaigned on pauses or tighter rules (TX, OH, GA, etc.). Local votes went heavily against new builds.
Markets reacted hard: Nasdaq -5.5%, SOX -12%+, $NVDA roughly -9-10% in one session. This wasn’t an AI-demand story. It was a “can we actually build it” story.
2/ The big hyperscalers won’t freeze CapEx. They also won’t keep charging ahead with the old playbook of tax breaks + NDAs + minimal community givebacks.
Here’s the more likely playbook:
3/ First move: buy peace with real money.
$META and $AMZN already floated ~$1B community packages. $GOOGL said new campuses will pay full property taxes. $MSFT has started walking away from some abatements. Expect bigger, more binding local deals — schools, job training, ratepayer protection funds — in exchange for permits.
4/ Second: geographic arbitrage.
If Texas stays sticky for 6-12+ months, capacity shifts toward states/regions with existing power, friendlier politics, or federal land options. The hyperscalers can move. Pure-play, high-leverage neo-clouds will feel more pain.
5/ Third: vertical integration on power.
More on-site generation, long-term power PPAs, and even temporary gas solutions. Higher upfront cost, lower political veto risk. The constraint is no longer just GPUs — it’s electrons + local consent.
6/ Fourth: CapEx guidance will get more cautious in the near term (especially 2028+ pipeline), but the long-term buildout continues. Competition between $MSFT, $GOOGL, $AMZN and $META is too intense to pause. They just have to build smarter and share more of the surplus.
7/ Bottom line:
“Expand first, negotiate later” still works — but the negotiation price just went up. The winners will be the ones who turn community resistance into locked-in local partnerships fastest while securing power. The ones who treat this as temporary noise will keep eating delays and higher costs.
AI isn’t going away any more than Tesla did when people couldn’t understand it — those who refuse to embrace it will simply be left behind by the planet, and because human nature is greedy, the data centers must be built first so capital has something real to negotiate over.
@aleabitoreddit The comment section is once again full of people who think they’re smarter than Musk, plus a few experts in the noble art of commercial plagiarism.
In what world is unrestricted short selling some kind of public service during a disorderly rout? Short sellers are not “providing liquidity” in any meaningful sense when the market is already in freefall. They are adding selling pressure first. The fact that they eventually have to buy back to cover does not magically make the net effect bullish or stabilizing. In practice, the initial wave of short selling often accelerates the decline, forces margin calls on longs, and creates the very panic the authorities are trying to stop.
Taiwan cutting borrowed-share short sales from 30% to 3% of volume, raising margin requirements, and standing up a stabilization fund is a standard crisis response, not some mysterious “bullish” magic trick. South Korea and China doing similar things points to the same logic: when leveraged speculative selling is amplifying an already sharp drop in AI/semiconductor names, temporarily restricting the ability to pile on more shorts can reduce forced selling and give the market a chance to find a floor. That is not the same as claiming the restrictions will permanently raise valuations.
Treating short sellers as uniquely virtuous liquidity providers while dismissing the downside risk they introduce is selective. A large share of short activity, especially in high-volatility sectors, is directional speculation that looks a lot like gambling with borrowed shares. Informed, fundamental shorting exists, but it is not what dominates when a sector is already collapsing and retail or momentum accounts are piling in on the short side. Authorities do not impose these rules because they are confused about market microstructure; they do it because unconstrained short selling has repeatedly turned volatility into a crash.
Your argument assumes that more short selling always improves outcomes because covers are buys. That only holds if the short selling did not create additional selling pressure, margin cascades, or panic in the first place. In the current episode, the evidence from the price action and the policy response suggests the opposite.
Trying to manipulate the narrative and stock prices on X is a quick way to catch an SEC investigation. Elon has been doing this kind of thing for years and somehow walked away completely unscathed recently. Meanwhile, the Democrats’ selective outrage and derangement have just reached absurd levels.
Every tech bro on your timeline thinks we’re 2 years away from *Detroit: Become Human*.
Then you look at the P&L of the “Embodied AI” pioneers, and you realize the real winner of the robotics revolution is… a toy factory in Guangdong printing fake plastic dogs. 🧵👇
---
**1/ Let’s look at Unitree’s latest Q3 2026 earnings.**
Hailed as the "First Pure-Play Humanoid Stock," the top line looks like a typical high-flying tech darling:
• 9M 2026 Revenue: ¥1.82B (~$250M), +55.7% YoY.
• Q3 Revenue: ¥665M, +70% YoY.
Bullish, right? Now run a basic DuPont analysis and check the recurring bottom line:
• 9M Core Net Profit (扣非净利): ¥349M, **DOWN -18.8% YoY**.
• Q3 Core Net Profit: ¥105M, **DOWN -17.4% YoY**.
• R&D Expenses: ¥263M, **UP +191.5% YoY**.
Revenue goes to the moon; earnings fall off a cliff. Why? Because selling cutting-edge robots is a brutal blender for shareholder equity.
---
**2/ The "Cyber-Charity" P&L**
Unitree claims profit took a hit because they are “heavily investing in the robotic brain and physical AI.”
In plain English: every dollar of margin squeezed out of their G1 humanoids and B2 quadruped dogs is immediately wire-transferred straight to Nvidia for GPU clusters, and to precision gear manufacturers for high-end actuators.
They aren’t running a high-margin tech flywheel; they’re running a non-profit subsidy pipeline for upstream hardware suppliers.
---
**3/ Stop projecting "Detroit: Become Human" onto your DCF models**
Retail traders price these stocks like Connor the Android is going to walk off the assembly line next Tuesday and replace half the blue-collar workforce.
Wake up. Look at Elon Musk.
Tesla has burnt billions of dollars in free cash flow (FCF) on Optimus. What do we have to show for it after years of hype? Staged demos, teleoperated arms serving drinks at parties, and a giant money-losing cost center. If the richest man on Earth with Tesla’s infinite manufacturing scale can’t make humanoids remotely cash-flow positive, what makes you think a 100x P/E pure-play can?
Real-world robots still suffer from sensor drift, battery limits, motor overheating, and a total lack of generalized zero-shot autonomy outside calibrated lab floors. We aren’t in *Detroit: Become Human*; we’re in *Roomba with extra joints*.
---
**4/ The ultimate humiliation: The "Fake Robot" Toymakers**
Now, contrast this with toy manufacturers and pop-culture brands (Pop Mart, Alpha Group, etc.):
• **Pop Mart (H1 2026)**: Made **¥5.16B (~$720M)** in adjusted net profit in just *six months*—that’s nearly **15 times** Unitree’s entire 9-month core net profit! Their plush monster doll (Labubu) alone raked in ¥18.7B in a single year.
• **Alpha Group (9M 2026)**: A legacy toy company selling plastic yo-yos and remote-controlled kiddie cars pulled in **¥2.06B in revenue**—literally beating Unitree’s ¥1.82B top line.
Let that sink in:
The company mastering dynamic balance algorithms and harmonic reducers is getting financially lapped by a factory making plastic cars for 8-year-olds.
---
**5/ The Trade Idea & Brutal Reality**
Toy factories don’t solve inverse dynamics. They don’t hire PhDs to train multimodal foundation models.
They inject ¥2 worth of cheap PVC into a mold, slap a $0.50 generic Bluetooth chip inside, print **"AI Smart Robot Companion"** on the packaging, and sell it to parents or nostalgic Gen Zs for a 70% gross margin. No firmware crashes, no battery warranties, zero customer support.
**The Bottom Line:**
Hardcore robotics engineers bet their balance sheets trying to build a *real* autonomous robot in the physical world.
Meanwhile, toy factory bosses on yachts figured out the ultimate market inefficiency:
**Humans will never pay for the true cost of real intelligence, but they will happily hand over billions in pure profit for a cheap, fake robot that makes them feel good.**
*Rating: LONG Plastic Toys, SHORT Cyber-Delusions.* 🧸📉
Just shipped QiLoom! 🎉 One family birthday shouldn’t need five tabs. Convert Gregorian/lunar dates, find lunar birthdays, and calculate baby milestones—all free. 📅 Try it: https://t.co/LO0aPxnsQI #BuildInPublic#IndieDev
Intel’s current crisis didn't happen overnight—it has been a relentless downward spiral ever since late 2020.Selling off its entire NAND business and Dalian fab to SK Hynix was an unforgivable strategic retreat, and completely alienating its core enthusiast user base by hiring radical feminist comedian Yang Li for an ad campaign in China was a self-inflicted PR disaster.From ceding ground to TSMC, butchering marketing, to botched capital allocation, Intel has taken every wrong turn possible. $INTC
Just shipped https://t.co/bhbI97Qr0J! 🎮🚀 Too many tabs for one game question? I built an independent hub for wikis, guides, builds, tier lists & codes. Find your game and follow the sources 👇 https://t.co/ZLfK0YxKxO #Gaming#BuildInPublic