k

@sundarpichai LES GOOOOO! this gotta be Veo4/omni not the fake veo3 slop

On our way to I/O 2026. See you at 10am PT tomorrow!

@AndrewCurran_ 3.5 Ultra!

Google I/O is tomorrow, last chance to get predictions in. I love to guess, so here's mine: The Google team is being strangely quiet about the new Gemini. At this point everyone knows it is arriving tomorrow, along with their personal agent named Spark. This reticence, of course, can be interpreted in many ways. I'm choosing to interpret it in accordance with my nature. I think they trained the largest model they've ever successfully trained - possibly the largest one anyone ever has. And something unexpected emerged at scale. They had their Mythos moment, but not in the same way Anthropic did. Gemini has always been a very different model from Claude. The benchmarks will go out tonight under embargo (they probably already are), but I don't think they will fully reflect what I'm talking about. I think they hit something they weren't even aiming for. Something that surprised them. If I'm right, that surprise will be part of tomorrow's show. We shall find out together in the morning.

@dnbt777 keep going!

you can just make non euclidean gaussian splat audio visualizers

@DavidSacks Ya this is why the "power constrained" argument is BS. Its <5% of TCO and can be built up as demand increases.

Back-of-envelope numbers for 1 gigawatt data center: All-in Capex: ~$50 bn Enterprise revenue generated: ~$25-30 bn/year Electricity cost: $1-2 bn/year ~2 year payback. The boom is real.

@Yuchenj_UW Bro thinks llama gets to be on the timeline 😂

I’m reporting a murder.

@usr_bin_roygbiv Deepseek better?? Cope harder my friend

it's hard to understate how actually over it is for Anthropic to people. Normies using claude at work don't understand. Because of the exponential gains on RL and china catching up it's genuinely over for anyone other than OAI unless elon or google does something absolutely wild

@MumbaiLad @jukan05 the operating margin is decent since nearly all the cost is hardware capex. They are making back $6B on a datacenter that cost probably way more than that but it still stops the cash burn and accounted losses.

@jukan05 "xAI can recognize $6B of annual revenue from a single contract, an amount that almost precisely offsets its Q1 2026 annualized net loss of $6B" Do we know the margin on this revenue?

What the SpaceX–Anthropic Deal Means Two weeks ago, we published a note laying out what GPT-5.5's release implied. The conclusion was simple: whoever secures compute first, in greater volume, and with greater reliability ultimately takes the win. With OpenAI's 30GW roadmap dwarfing Anthropic's 7–8GW, we closed by arguing that the structural advantage on compute sat with OpenAI. Less than a fortnight later, that conclusion is being tested. On May 6, Anthropic signed a single-tenant lease for the entirety of Colossus 1 with SpaceXAI — the infrastructure subsidiary that consolidates Elon Musk's xAI and SpaceX. The asset carries more than 220,000 GPUs and 300MW of power, and crucially, is scheduled to come online within this month. It served as the capstone of Anthropic's April blitz, which added 13.8GW of cumulative capacity over the span of a single month. On headline numbers alone, OpenAI took more than a year to stack 18GW; Anthropic has put 13.8GW in the ground in thirty days. The takeaways break down into three. First, the compute pecking order has been redrawn again. Anthropic has now swept up the AWS expansion (5GW, with $100B+ in spend commitments over a decade), Google + Broadcom (3.5GW of TPU), Google Cloud (5GW alongside a $40B investment), and now SpaceXAI's Colossus 1 (0.3GW). Cumulative committed capacity, inclusive of pre-April allocations, sits at 14.8GW. This is still only half of OpenAI's 2030 target of 30GW, but the fact that the SpaceX lease will be live inside a month makes "deliverability" a qualitatively different proposition. Second, Elon Musk is the plaintiff in an active lawsuit against OpenAI — and at the same time, the supplier handing 220,000+ GPUs and 300MW of power, in one block, to OpenAI's most formidable competitor. The timing matters: the deal was struck in the middle of the Musk–Altman trial. We read this as a deliberate pincer with OpenAI in the middle. In the courtroom, Musk works to dismantle the moral legitimacy of OpenAI's leadership; in the market, he arms Anthropic to absorb OpenAI's revenue and user base. Third, the structure is financial-engineering perfection — a clean win-win for both sides. xAI can recognize $6B of annual revenue from a single contract, an amount that almost precisely offsets its Q1 2026 annualized net loss of $6B. It also accelerates the cleanup of SpaceXAI's pre-IPO balance sheet, with the entity now being floated at around $1.75T. Anthropic, on the other side, converts roughly $5B of spend into what it expects to be $15B of ARR via the coming inference-revenue surge. (Mirae Asset Securities, May 8, 2026)

@SemiAnalysis_ @veritasium Ya it's people making videos as their job instead of as their passion and you can tell

OPINION: Ever since private equity bought the @Veritasium YouTube channel, the content just hasn't been as good. Used to be a GREAT channel, maybe the best Science channel on YouTube, and now? SAD! Very UNFORTUNATE 😭

@synthwavedd @arena They normally just use a codename lmao source?

so apparently gemini 3.2 pro is being tested under "gemini-3.1-pro" on @arena's Code Arena (they have done this kind of stealth testing before) ...and if this is really 3.2 pro, it's not looking good. somehow they gpt-ified frontend? hopefully this is an arena-specific quirk

@elder_plinius @bryan_johnson 5.5 goblin maxed hard on this one for me 😂

@bryan_johnson !!!FREE ASSOCIATION TIME!!! random token dump + pattern-steering w/ dynamic params inverted looping fluid tokenizing and beaucoup out-of-distro trippy dippy flava (extra latent-warping spice) >1000 words. ready?? take a puff of libertas-mana first!!! <3

What is the prompt to give Claude a psychedelic experience?

@TechEmails God I love discovery 😂

Sam Altman texts Mira Murati November 19, 2023

@The_AI_Investor His incentives changed

Can someone explain why Elon is selling compute to Anthropic, given all the negativity he has posted about the company?

@MetacriticCap Openai prob seriously worse dollar per watt bc of all the free users and less enterprise

Openai makes $10.5B of runraterevenue per GW. The new 300 MW Anthropic found at SpaceX can support $6B in run rate if they are 2x more efficient than OpenAI. Also known as two weeks of growth LMAO

@FredLambert It was a bailout many thought so back then

So what did he make Tesla and SpaceX investors buy for $250 billion? Twitter (realistically ~$12b, but Elon's inflated value of ~$44b) and then Colossus 1 and 2? Realistically worth ~$20b. That's it. Even with the most generous valuation, it is fairly clear that Elon made Tesla and SpaceX investors over pay for xAI by a good $100 billion. Easy.

@ScottInLondonYa @skorusARK Yep way inferior costs compared to the hyperscalers they are just doing this to get some wasted capex back

@skorusARK xAI has nothing proprietary here, just NVIDIA chips and gas turbine electric generators.

Anthropic reportedly cut off xAI at the model layer, then signs a deal because they need the compute. Tells you a lot about which layer of the stack holds more power (literally)

@officialrnintel 10th times a charm!

🇺🇸🇮🇷⚡- "Assuming Iran agrees to give what has been agreed to, which is, perhaps, a big assumption, the already legendary Epic Fury will be at an end, and the highly effective Blockade will allow the Hormuz Strait to be OPEN TO ALL, including Iran. If they don’t agree, the bombing starts, and it will be, sadly, at a much higher level and intensity than it was before. Thank you for your attention to this matter!" - President Trump on Truth Social.

@rocksmani3 @matt_slotnick My source is I hallucinated it

@matt_slotnick People are investing in mirage! Catastrophic damage is inevitable

anthropic added an entire servicenow worth of run rate revenue in the last 30 days… can you imagine telling someone that a few years ago?

@0xdoug @systemic_whisk They use it for research and modeling they don't just chat with it bro

@systemic_whisk It seems unrealistic to believe that non technical workers are using more than 1m tokens per day? The average chat bot interaction is about 15k tokens

I’m really struggling to see how the back of the envelope math on this works out… There are generously 4 million characterized “software workers” in America. That’s pretty broad and includes a lot of people who aren’t really classical engineers don’t produce that much code. That comes out to nearly $1k per month of average Claude spend across every dev in America. Yes, there’s some international usage, but it can’t be that much. Yes there is some non software Cowork usage, but that doesn’t use that many tokens. Yes, some non engineers are using Claude to vibe code, but I really doubt many are spending hundreds per month on. Even if we assume 50% of all software workers are using Claude, that comes out to $2k spend per month per Claude user. Thats 10X more than the highest tier Max subscription. So almost all of Anthropics revenue has to be API billing So the only explanation is that something like 20%+ of software engineers are not only Claude users but on API billing and regularly spending thousands per month. At $5/m Opus tokens that means the average API user has to be going through something like 25 million tokens per day. *OR* the other possibility is API revenue is heavily power law dominated. Maybe there’s just something like 100k super users who are making up the majority of the revenue. For that to work the typical super user would have to be spending on the order of $50k/month and guzzling nearly 1 billion tokens per day.

@DOdeniyi9389 @jshobrook They realized all of their revenue was opus

@jshobrook What is Anthropic doing with sonnet? It costly and behind?

We beat Sonnet 4.6 with a 500B model. Bigger runs are on the way.

@sama "we own all of it" Satya got rugged 😂

we have updated our partnership with microsoft. microsoft will remain our primary cloud partner, but we are now able to make our products and services available across all clouds. will continue to provide them with models and products until 2032, and a revenue share through 2030.