bare metal (@embvded) header
bare metal (@embvded) profile picture

c++ is the only language you need. learning to create art. get good.

Joined November 2019
Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)Photo by bare metal (@embvded)
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…
176
147
32
1,955
611,754
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.
46
436
53
1,313
24,529
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🧵
90
238
61
2,888
299,992
“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.
1,060
19,557
1,078
61,274
1,504,762
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.
19
34
6
199
11,756
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…
27
29
4
308
41,779
bare metal retweeted
Replying to @elonmusk
My attempt at drawing elon. Day 16 of drawing elon for a free supergrok subscription
1
1
1
34
bare metal retweeted
Replying to @elonmusk
A bit of a learning curve. Day 15 of drawing elon for a free supergrok subscription
1
1
50
bare metal retweeted
Replying to @elonmusk
Day 14 of drawing elon for a free super grok subscription
1
2
93
bare metal retweeted
Replying to @elonmusk
Drawing elon for a free supergrok subscription
1
179
bare metal retweeted
Replying to @elonmusk
Day 13 of drawing elon for a free supergrok subscription will get good
1
17
bare metal retweeted
@grok pick a random follower that Retweet & share this post for another $100 from my payout in 10hrs
230
326
29
601
21,901
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.
1,979
4,609
4,525
42,681
9,062,633
bare metal retweeted
Replying to @elonmusk
Reading human all too human now
1
408
bare metal retweeted
Replying to @elonmusk @bot
Notable descent in capabilities💀 it doesn't even look like elon now, I'll be better tomorrow. Day 12 of drawing elon for a free supergrok subscription
1
1
1
526
bare metal retweeted
imagine this happens lmao
125
123
35
6,107
131,082
bare metal retweeted
Introducing Eleven v4 and Eleven v4 Turbo, our fastest and most emotive voice models yet. Ranked #1 by Artificial Analysis.
517
1,530
1,160
19,248
5,013,674
bare metal retweeted
Replying to @elonmusk
Day 9 of drawing elon for a free supergrok subscription
2
1
46
bare metal retweeted
Replying to @elonmusk
Day 8 of drawing elon for a free supergrok subscription
1
59
bare metal retweeted
Replying to @elonmusk
Day 7 of drawing elon for a free supergrok sub
1
36