bare metal retweeted
X premium free for a month, thank me later!
Expires on: Monday, November 30, 2026 at 10:30:00 PM
Redeem At: grok.com/plans/referral/?sku…
Readers added context they thought people might want to know
The post's claim of "X premium free for a month" via referral is inaccurate. Official program offers eligible new users a 14-day trial requiring a credit card at checkout with auto-renewal unless cancelled; referrers must meet US eligibility rules and receive usage resets.
x.ai/legal/referral…
docs.x.ai/x-subscription…
bare metal retweeted
“Islamophobia” If you're not phobic from this, there's something genuinly dark and twisted inside you and you're an extremely evil person lacking of sympathy.
bare metal retweeted
I improved NanoGPT Speedrun from 39.9s → 6.7s (-83.2%), more than every record since 2025 combined (-81.4%)!
The web is full of rare, repeated content that the LM can’t memorize in time. I solve this with retrieval-augmented n-gram hints🧵
bare metal retweeted
“She is a sister. Don’t loot here, let’s go there!”
Muslim rioters in France recognize the shop owner as a fellow Muslim just before nearly destroying her store by accident.
Not a single Muslim-owned business or mosque has been looted in the riots.
bare metal retweeted
While QAT helps improve model quality at 2-4 bits, it has a structural optimization problem. Indeed, at every step, the high-precision latent weights receive the gradient updates, but the loss and gradients are computed on their low-bit reconstructions, i.e. the weights the model will use at inference.
This mismatch results in suboptimal training dynamics and a higher loss floor. The main existing QAT methods (learning the scales alongside model parameters, soft-to-hard rounding, or better initialization) do not directly address this mismatch.
With QUASAR, we introduce a simple and computationally cheap way to address the mismatch. QUASAR consistently achieves lower loss than all QAT baselines and adds no inference overhead. Also, it can be combined with most other QAT methods, since it only changes the weight reconstruction map in the fwd pass.
bare metal retweeted
Congrats to @hylight_aero (S23) on their $14.5M raise!
They’re bringing back the airship, this time autonomous, hydrogen-powered, and built to inspect critical infrastructure.
Power lines, pipelines, and railways are still mostly inspected by helicopter, which is expensive, polluting, and risky for crews. HyLight’s 12-meter airships hover low and slow with LiDAR, cameras, and methane detectors, capturing imagery down to 1 millimeter per pixel. Four are flying missions in France today, with 15 planned to operate at once by 2027.
aerotime.aero/articles/frenc…
bare metal retweeted
Replying to @elonmusk
My attempt at drawing elon. Day 16 of drawing elon for a free supergrok subscription
bare metal retweeted
Replying to @elonmusk
A bit of a learning curve. Day 15 of drawing elon for a free supergrok subscription
bare metal retweeted
Replying to @elonmusk
Day 13 of drawing elon for a free supergrok subscription will get good
bare metal retweeted
Introducing Gemini 4 Argon – our new frontier model.
It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
bare metal retweeted
Introducing Eleven v4 and Eleven v4 Turbo, our fastest and most emotive voice models yet.
Ranked #1 by Artificial Analysis.





















