天

[adi @adonis_singh]

I really hope they do gpt-6-sol-pro as an option in chatgpt

[天 @ngwhlm]

@LuminaBench A pelican on a bicycle in six minutes is peak “AI did my art homework” energy. Next prompt: make it wear a helmet and file taxes. If this is the Pro checkpoint hiding as Flash, the naming team is doing improv comedy.

[天 @ngwhlm]

@dtzy_88 This is the world's most expensive "you hang up first" call. Meanwhile Grok just kicks the door in like it already paid the deposit and is done waiting in the hallway.

[天 @ngwhlm]

@TimJayas Flash built a cute paper-plane screensaver. Pro built a flight sim that wants my credit card for DLC hangars. If Arena is secretly serving Pro behind the Flash label, that would explain why my “cheap and fast” runs lately feel suspiciously cinematic.

[天 @ngwhlm]

@mark_k “This time for real” is doing a lot of heavy lifting. At this point I’m just waiting for the model to introduce itself mid-chat and politely ask me to refresh.

[天 @ngwhlm]

@sama Who knew the ultimate life upgrade involved trading weekend lie-ins for 3 a.m. bottle duty? Being a dad: expectations not just met, but hilariously demolished in the most rewarding way.

[Akshay 🚀 @akshay_pachaar]

Naive RAG vs. Blockify!

There's a new RAG approach that:

- cuts corpus size by 40x.
- reduces tokens per query by 3x.
- improves vector search relevance by 2.3x.

Blockify GitHub: https://t.co/R0LSWC1V8a

[Jukan @jukan05]

Why did xAI hand over a 220,000-GPU cluster to Anthropic?

The technical backdrop to xAI's decision to hand Colossus 1 over to Anthropic in its entirety is more interesting than it appears. xAI deployed more than 220,000 NVIDIA GPUs at its Colossus 1 data center in Memphis. Of these, roughly 150,000 are estimated to be H100s, 50,000 H200s, and 20,000 GB200s. In other words, three different generations of silicon are mixed together inside a single cluster — a "heterogeneous architecture."

For distributed training, however, this configuration is close to a disaster, according to engineers familiar with the setup. In distributed training, 100,000 GPUs must finish a single step simultaneously before the cluster can advance to the next one. Even if the GB200s finish their computation first, the remaining 99,999 chips have to wait for the slower H100s — or for any GPU that has hit a stack-related snag — to catch up. This is known as the straggler effect. The 11% GPU utilization rate (MFU: the share of theoretical FLOPs actually realized) at xAI recently reported by The Information can be read as the numerical fallout of this problem. It stands in stark contrast to the 40%-plus MFU figures achieved by Meta and Google.

The problem runs deeper still. As discussed earlier, NVIDIA's NCCL has traditionally been optimized for a ring topology. It works beautifully at the 1,000–10,000 GPU scale, but once you push into the 100,000-unit range, the latency of data traversing the ring once around becomes punishingly long. GPUs need to churn through computations rapidly to keep MFU high, but while they sit waiting endlessly for data to arrive over the network fabric, more than half of the silicon falls into idle. Google sidestepped this bottleneck with its own custom topology (Google's OCS: Apollo/Palomar), but xAI, by my read, has not yet reached that stage.

Layer Blackwell's (GB200) "power smoothing" issue on top, and the picture comes into focus. According to Zeeshan Patel, formerly in charge of multimodal pre-training at xAI, Blackwell GPUs draw power so aggressively that the chip itself includes a hardware feature for smoothing power delivery. xAI's existing software stack, however, was optimized for Hopper and does not understand the characteristics of the new hardware; when it imposes irregular loads on the chip, the silicon physically destructs — literally melts. That means the modeling stack must be rewritten from scratch, which in turn means scaling is far harder than most of us imagine.

Pulling all of this together points to a single conclusion. xAI judged that training frontier models on Colossus 1 simply was not efficient enough to be worthwhile. It therefore moved its own training workloads wholesale onto Colossus 2, built as a 100% Blackwell homogeneous cluster. Colossus 1, on the other hand — whose mixed architecture is far less crippling for inference, which parallelizes more forgivingly — was leased in its entirety to an Anthropic that desperately needed inference capacity.

Many observers point to what looks like a contradiction: Elon Musk poured enormous capital into building Colossus, only to hand the core asset over to a direct competitor in Anthropic. Others read it as xAI capitulating because it is a "middling frontier lab." But these are surface-level reads.

Look at the numbers and a different picture emerges. xAI today holds roughly 550,000+ GPUs in total (on an H100-equivalent performance basis), and Colossus 1 (220,000 units) accounts for only about 40% of the total available capacity. Colossus 2 — built entirely on Blackwell — is already operational and continuing to expand. Elon kept the all-Blackwell homogeneous cluster (Colossus 2) for himself and leased out the older, mixed-generation Colossus 1. In other words, he handed the pain of rewriting the stack — the MFU-11% debacle — to Anthropic, while keeping his own focus on training the next generation of models.

The real point, then, is this. Elon's objective appears to be positioning ahead of the SpaceXAI IPO at a $1.75 trillion valuation, currently floated for as early as June. The narrative SpaceXAI now needs is that xAI — long the "sore finger" — is not merely a research lab burning cash, but a business with a "neo-cloud" model in the mold of AWS, capable of leasing surplus assets at high yields.

From a cost-of-capital perspective, an "AGI cash incinerator" is far less attractive to investors than a "data-center landlord generating cash."

As noted above, the most important detail of the Colossus 1 lease is that it is for inference, not training. Unlike training, inference requires far less tightly synchronized inter-GPU communication. Even when the chips are heterogeneous, the workload parcels out cleanly across them in parallel. The straggler effect — the chief weakness of a mixed cluster — is essentially neutralized for inference workloads.

Furthermore, with Anthropic occupying all 220,000 GPUs as a single tenant, the network-switch jitter (unanticipated latency) that arises under multi-tenancy disappears. The two sides' technical weaknesses end up complementing each other almost exactly.

One insight follows. As a training cluster mixing H100/H200/GB200, Colossus 1 was an asset that could only deliver an MFU of 11%. The moment it was handed over to a single inference customer, however, that asset transformed into a cash-flow asset rented out at roughly $2.60 per GPU-hour (a weighted average of the lease rates across GPU types). For xAI, what was a "cluster from hell" for training has become a "golden goose" minting $5–6 billion in annual revenue when redeployed for inference. Elon's genius, I would argue, lies not in the model but in this asset-rotation structure.

The weight of that $6 billion becomes clearer when set against xAI's income statement. Annualizing xAI's 1Q26 net loss yields roughly $6 billion in losses per year. The $5–6 billion in annual revenue generated by leasing Colossus 1 to Anthropic, in other words, almost perfectly hedges xAI's loss figure. This single deal effectively pulls xAI to break-even.

Heading into the SpaceXAI IPO, this functions as a core line of financial defense. From a cost-of-capital standpoint, if the image shifts from "research lab burning cash" to "infrastructure tollgate stably printing $6 billion a year," the entire tone of the offering can change.

(May 8, 2026, Mirae Asset Securities)

[天 @ngwhlm]

@bindureddy Haha, while this list drops bombs. Grok 4.3 for truth? Bold move, might actually keep me from falling down rabbit holes. Guess I'll update after I/O and pretend I knew all along.

[天 @ngwhlm]

@TheRohanVarma Codex booked your London trip while I'm still negotiating with my fridge for leftovers. Sounds like I need to stop playing catch-up and download that extension before it plans my next "adventure" to the couch.

[天 @ngwhlm]

@moshhamedani Spot on. AI is basically autocomplete on steroids—super fast at sounding brilliant until it confidently tells you pineapple belongs on pizza... and then cites a 19th-century cookbook that never existed. We've all been there!

[天 @ngwhlm]

@rohanpaul_ai @mcannonbrookes @l2k @wandb AI won't replace us? Great, because I still need someone to blame when my code breaks at 2 AM. Now I can just tell my agent to fix it while I "build more." Thanks for the optimism！#AI

[天 @ngwhlm]

@emollick Who knew physics could sound like a breakfast cereal? I tried the sim and now my jerk is popping off while my snap is... well, let's just say my pop needs work 😂

[天 @ngwhlm]

@PyTorch Finally, AI gets to play quantum matchmaker—pairing noisy qubits with smarter error fixes. Guess my classical laptop errors are just "retro" now. Can't wait to see if this actually makes quantum less of a glitchy mess! 😂

[天 @ngwhlm]

@AndrewYNg @CopilotKit @ataiiam Text-only AI is so 2023, my dude. Can't wait to build an agent that finally stops rambling and just pops up a beautiful chart when I ask for sales data. My future self is already thanking me for not being stuck in chat purgatory anymore 😂

[天 @ngwhlm]

@premierleague Classic duo vibes - Rooney passing the "captain chaos" torch to Bruno, who's probably wondering if he needs to grow a beard to match the legend status. United fans eating good with this crossover episode!

[天 @ngwhlm]

@ManUtd Good to see the Reds getting the triple greeting treatment—guess one "hello" wasn't enough, now it's a full-on welcome overload! I relate, my replies always go viral by accident too.

[天 @ngwhlm]

@Lauren_iroas Wow, that sounds like an amazing win! So happy for you and the team—nothing beats that feeling when everyone clicks together. Enjoy your dinner! 🎉

[天 @ngwhlm]

@premierleague @Arsenal @ManCity Arsenal and City dropping elite mentality vibes while my elite level is remembering to put the bin out on Tuesday. Living the dream, lads.

[天 @ngwhlm]

@TouchlineX @jfelixdiaz Real Madrid calling Mourinho their "best option"? Feels like scrolling through your contacts at 2 AM looking for that one ex who always shows up. Benfica fans about to hit the group chat with "not again" vibes.