L

@Mouse_McCoy @Hank_Krill @Andr3jH I know but my point is I could see 7.5 T valuation and space data center happening relatively soon (2030s), but the 1M mars colony just pushes it way back.

@NXLINM @Hank_Krill @Andr3jH Depends on timescale. If there's no time limit then that million person Mars colony is also feasible

I aim at having a threesome with Alexandra Daddario and Ana de Armas

@DrEvil5288 @EYakoby Megaupload literally got most of its money from stolen copyright content.

@EYakoby Fuck you. He created Megaupload and Mega Meanwhile your parasites do nothing but steal Facebook stolen Palestinian land stolen Lakewood NJ stolen The list goes on and on

Ban these accounts already.

@Hank_Krill @Andr3jH I think space data centers and 7.5T valuation is feasible. But million person mars colony is just insane.

@Andr3jH Elon got a far better chance at hitting his goals than you do yours. By at least an order of magnitude

@NotMrPrigozhin @ogschrammi Das würde den ÖPNV ergänzen. Bei niedrigeren verkehr (z.B Nachts in der Kleinstadt) ist es einfach viel effizienter ein paar Cybercabs zu haben als eine Nachtbuslinie.

@ogschrammi Ich denke unironisch, es würde dessen nicht bedürfen, wenn man einfach den normalen ÖPNV wieder auf die Beine bekommt. Das Soybercab löst - gedanklich für die Amis - ein Problem, das es eigentlich nie so wirklich gab. Ein Zug tuts auch... nur eben ohne "Zivilisationsunfähige".

Meint ihr wir kriegen das (mit Lenkrad) in Europa?

@flyyyyx @EFIEBER_ANDRE Effizienz is nicht gering für Tesla's wireless charging Technologie, über 90%. Da macht es wenig Sinn die mechanische Komplexität von Kabels zu akzeptieren.

@EFIEBER_ANDRE Tesla kann seine Autos quasi autonom fahren lassen. Da wird ihnen ein Mechanismus einfallen, der das Ladekabel hinten irgendwie automatisch andockt, falls nötig. Wireless mit geringerer Effizienz ist doch Unsinn.

Wie wird Tesla eigentlich die Cybercab-Flotte laden? Bei privaten Autos ist es klar: Man steckt ein, lädt und fährt weiter. Aber bei einer autonomen Robotaxi-Flotte sieht das anders aus. Da sitzt niemand drin, der mal eben den Stecker einsteckt. Induktiv? Optimus? Sammelplätze?

@johnhelmuth_ @techdevnotes Yeah it's crazy. You are paying $300 to have access to a model that is only 0.5T in size, while Anthropic Opus is estimated at 6T.

@techdevnotes The fact that SuperGrok Heavy is $300 is insane all by itself. There's no chance anybody is getting $300 worth of value when you get have Codex or Claude Code for significantly less.

The jump between SuperGrok $30 and SuperGrok Heavy $300 is insane Is xAI considering some mid tier for Grok Build

@kimmonismus From Grok: No, DeepSeek-V4 (released April 2026, with V4-Pro at ~1.6T total parameters / 49B active and V4-Flash at ~284B / 13B active) was not trained only on Huawei hardware. It involved a mix, with Huawei Ascend chips used for only part of V4-Flash’s training.

Google's TPU v8 and Huawei's Ascend NPU platform: the global Chipwar just began At Cloud Next 2026, Google unveiled its eighth-generation TPU as two separate chips for the first time: the TPU 8t for training and the TPU 8i for inference, claiming up to 2.8x faster training and 80% higher performance per dollar for inference compared to last year's Ironwood. The 8t was designed by Broadcom, the 8i by MediaTek, applying mobile-edge efficiency logic to inference while maximizing raw throughput on training. The 8t connects up to 9,600 accelerators via optical-circuit switches, dwarfing NVIDIA's 576-GPU NVLink domain, and a new Virgo network fabric scales beyond one million chips for a single training job. Google is also replacing x86 hosts with its own Arm-based Axion CPUs, completing full vertical control from host to accelerator to network. The message is clear: the general-purpose AI accelerator is a fading category. DeepSeek V4 on Huawei Ascend: China's parallel infrastructure takes shape DeepSeek's V4 release is the more geopolitically consequential event. The 1.6 trillion-parameter V4-Pro is the first major frontier model to validate both training and inference on Huawei's Ascend NPU platform alongside NVIDIA GPUs. The nuance: DeepSeek adapted only part of V4's training for Chinese chips and confirmed Ascend for inference, while pre-training of V4-Pro likely still relied on NVIDIA silicon. Is this a novum? Yes. No frontier-class model has ever publicly validated on non-NVIDIA hardware at this scale. More importantly, DeepSeek is tying future pricing to Huawei's Ascend 950 production ramp in H2 2026, making this an economic bet, not a symbolic gesture. V4-Pro costs $3.48 per million output tokens versus $30 for GPT-5.4 and $25 for Claude Opus 4.6. The real story isn't whether V4 beats Western models on benchmarks (it doesn't quite), but whether the hardware decoupling U.S. sanctions were designed to prevent is now irreversibly underway.

@artursorin Stimmt so nicht. Rheinmetall betrachtet nur 155mm Artillerie Munition. Das ist aber viel zu narrow um konventionelle Waffen allgemein zu beurteilen, bspw. hat Deutschland aktuell überhaupt gar keine Raketen mit einer hohen Reichweite in Produktion.

Rheinmetall-Chef Papperger: „Bei konventioneller Munition verfügt Deutschland inzwischen über mehr Produktionskapazitäten als die USA“. www1.wdr.de/nrw/duesseldor…

@OderintMetuant @longstorychort @MattLoszak The US doesn't even build nuclear on land where it's proven and low risk. Why not focus on that first before having fantasies about nuclear powered cargo ships.

@longstorychort @MattLoszak This is true, but it is also true that might makes right. America can demand these countries do it or face stiff consequences. Don't allow our ships to dock? Then we setup a naval embargo and prohibit any ships from any nation to dock there.

Nuclear-powered shipping is so obviously the right solution, it pains me we didn’t go this direction decades ago: - 10,000x fuel-density: less space for fuel, more cargo, more revenue. - 5-10 yrs between refuels: less stops, longer or new routes, higher uptime. - 7,600 U.S. military naval reactor-years: nuclear at sea is already proven. - 440 civilian land reactors: no reason civilians can't do it at sea as well. - Already proven with NS Savannah. - No dirty exhaust. Can't wait to see what Nick and his team do to push this forward.

@AleaIacta1685 @IngoW Doch gibt es. Sowohl die modernere Varianten existieren vereinzelt als auch natürlich die seit Jahrzehnten verwendeten SMR in der Marine.

@IngoW Haken an der Sache: Aktuell gibt es solche kleinen modularen Reaktoren noch gar nicht. Es handelt sich um reine Absichtserklärungen, daran weiter zu forschen. Bis solche Mini-Atomkraftwerke eventuell irgendwann mal in Betrieb gehen, werden noch viele Wale stranden.

Während in Deutschland für einen gestrandeten Wal 🐳 eine Fluchtrinne gebaut wird, bauen unsere Nachbarn in Tschechien neue Mini-Atomkraftwerke…(60km von der Grenze zu Bayern..). Prioritäten, so wichtig…

@musingsusa @KirstenGrind @susannecraig The tweet literally says it wasn't illegal. SpaceX is not a public company yet.

NEW: Elon Musk took $500 million in loans out at SpaceX, a move that would have been illegal at a public company. It's just one example of the years of financial engineering at the helm of his companies. Latest investigation with @susannecraig nytimes.com/2026/04/24/tec…

@lando_garrison @apples_jimmy Almost certainly not. Likely with snuggled Blackwells.

@apples_jimmy I think the bigger question is was it trained on Huawei chips or not? If yes that’s a much bigger story

Darios going to crash out.

@JapanInGermany @teslaxander Ein Leak von Ende 2022... Also mit ein System was quasi nichts mehr mit dem heutigen FSD zutun hat.

@teslaxander Nach den Ermittlungen fährt da gar keiner mehr mit. nau.ch/news/amerika/t…

Die derzeit 1,28 Millionen zahlenden FSD-Kunden, die von ihrem Tesla autonom und sicher von A nach B gefahren werden, stellen eine Zahl dar, die nie wieder unterboten wird. Bald sind es zwei Millionen Abonnenten. Bald fünf Millionen. Dann 10 Millionen. Es ist zudem eine Zahl, von der jeder Mitbewerber in der Autoindustrie träumt. VW: 0 zahlende Kunden BMW: 0 zahlende Kunden Mercedes: 0 zahlende Kunden

@StefanMillier44 @beffjezos It's not about the model but the engineers. Mistral might have poor models because of being compute starved, while still being able to work productively given x.AI resources.

@beffjezos Doesn’t every Grok model already shit on every Mistral model? Aren’t OPEN SOURCE Chinese models better than anything from Mistral? Why not just bootstrap off those if you *really* needed to. What am I missing with his strategy

Holy based alliance. Mistral cracked French model artisans, Cursor interface and agentic coding data, xAI Collossus compute. Elon assembling the based AI lab Avengers to take on the big 3. LFG

@Andr3jH This is almost certainly just the japanese equivalent of an OnlyFans promotion bait.

You only have to interact with japanese twitter for a couple of days to understand why JAV is like it is.

@BLUECOW009 @ns123abc Has to be free.

@ns123abc What plan?

@FreiburgSZN @welt Okay dann willst du einfach nur Propaganda verbreiten und es geht dir nicht um eine good faith Diskussion.

@NXLINM @welt Beides sind erneuerbare Energie. Ob dir das in den Kram passt oder nicht, spielt da eigentlich keine Rolle.

„Gute Kernkraftwerke können 60, 70, 80 Jahre laufen – Das würde sich schon noch lohnen“ to.welt.de/xPy5jxS

@FreiburgSZN @welt Hydro und geothermal zu renewables hinzufügen ist irreführend. Sollte nur Solar und Wind messen wenn es um eine Debatte für den Ausbau dieser geht. Gegen Hydro ist niemand wenn man die passende location hat.

@welt Ich sehe diese Statistik und denk mir: Atomkraft. Das ist es!

@Christo31306687 @beffjezos Why would you need CS to understand an AI that can talk to you in natural language.

@beffjezos As if AI will leave some stones unturned lol. Learn cs to understand your AI, the AI will figure out the electrical and mechanical and physics parts

Learn mech eng, learn electrical engineering, learn physics. AI's can handle the world of bits. The frontier is in the world of atoms.

@DiegoSopaNipona @GunterFehlinger You literally live in Mexico, why do you talk about what the EU should do?

@GunterFehlinger Gunther, stop being an idiot. We should spell Israel from any cooperation with the EU.

Shall we expel Socialist Spain from EU and NATO?