Robin (@rodoume) header
Robin (@rodoume) profile picture

I do data and machine learning stuff. Lost in Engineering Management @rodoume@sigmoid.social

Paris
Joined July 2010
Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)Photo by Robin (@rodoume)
Pendant que l IA a peut être résolu l hypothèse de Riemann aux États-Unis, en France il paraît qu'elle sait pas faire de bibliographie.
24
73
2
1,011
21,156
Le génie de GTA, c'est de laisser les jeunes générations vivre des rêves impossibles comme voler un avion de ligne ou accéder à la propriété.
Bir yayıncı, GTA 6'daki ilk 45 dakikasını evine tadilat yaparak geçirdi.
157
14,709
123
212,018
1,973,940
Robin retweeted
if i had a billionaire husband and no prenup, i’d introduce him to the Chinese spy myself
JUST IN: Jensen Huang reportedly has no prenup, meaning a divorce could cost him $81 billion, the most expensive divorce ever.
655
11,780
411
178,474
3,183,265
Replying to @rough__sea
It’s software engineering not software typing
30
225
29
3,451
296,600
Robin retweeted
JENSEN! THATS NOT WHAT WE MEANT BY CHINESE MODELS!!!!!!!
591
8,119
701
122,635
2,698,681
Next time you think that short sellers just make shit up - have a look at this SIX MONTH OLD tweet - which basically alleges the affair that is now news.
Replying to @CulperResearch
9) $NVDA CEO Huang has walked alongside Megaspeed at every step, and seems to have a close relationship with Megaspeed's Alice Huang. They partied together in June 2024 - same month as the Apex/Alibaba pledge, then dined together in May 2025 - just after U.S. trade restrictions.
85
709
28
8,116
848,728
Telling my wife I’m having an affair with a 26 year old Chinese spy for national security reasons
Everyone saying “Oh Jensen got got. Jensen fell for the honeypot” Ever consider he knew the whole time? That actually HE was honeydicking China and feeding them false intel? It’s a new Cold War. There’s levels to this. Don’t jump to conclusions. Don’t rule out a honeydick
40
144
9
5,299
171,574
Another unexpected comparison: one of our data centers uses about as much water in a year as a single Krispy Kreme store. Not the whole chain. One store.
BREAKING: Krispy Kreme officially outperforms NVIDIA over the past year.
382
781
124
17,138
1,303,756
Every VP needs their 30 pixels
Anthropic: We have the best model. Google:
8
26
1,937
66,483
Robin retweeted
The man is a prophet martin.kleppmann.com/2025/12…
3
16
3
280
45,886
We're in a permanent state of "Anthropic is ahead, OpenAI is right behind them, and Google has announced a model that's ahead but is actually way behind."
Google just announced new SOTA model team just announced gemini 4 carbon. their new best model and numbers are looking good > beats opus 5.5 > 65.2% on terminal-bench 4.0 > $2 input / $10 output per 1M tokens but again, it’s not available to everyone right now. only small group of cyber defenders will get access first google gotta stop doing this and release model asap if someone from google is watching this, please gib access. i wanna test it 😭
16
7
3
270
108,309
Genuinely funny from FT
23
533
52
8,468
257,225
anthropic's Frontier Red Team has gone TWO YEARS without ever abliterating a single model until GLM-5.3 two weeks ago
24
23
7
1,999
149,587
78 years ago CS Lewis wrote down exactly what my thoughts are on AI.
Basically what CS Lewis said about the threat of atomic bombs
3
1
55
3,686
Remember that Anthropic (along with ~all other agentic solutions) need this to be ALREADY DONE before using agents on your data
Claude Dashboards and Claude Motion are in beta today. Ask Claude to turn your data into live dashboards and your ideas into animated explainers.
7
22
3
230
18,456
Robin retweeted
Of course, that’s your contention. You’re a first-year machine-learning engineer. You just got finished readin' Attention Is All You Need, probably watched a Karpathy video too. So now you’re convinced everything is just transformers and scaling laws. That's gonna last until next month when you discover convolution, and then you're gonna be talkin' about inductive biases and locality and how CNNs were actually incredibly compute-efficient for vision. Then you're gonna read the FlashAttention paper, and suddenly everything's about IO complexity and SRAM and how FLOPs don't matter because the whole goddamn thing is memory-bandwidth bound. That'll last until somebody shows you an MoE model, and then you're gonna be in here regurgitating DeepSeek, talkin' about expert parallelism and active parameters and how dense models are economically obsolete. Then six months from now you'll write one shitty Triton kernel, look at an Nsight trace for the first time, and start telling everybody Python isn't actually the bottleneck because the GPU is asynchronous. And by next year you're gonna be standing right here explaining to me that your 400-billion-parameter model is fast because you turned on CUDA graphs.
Ben Affleck reveals he writes Python, understands convolutional neural networks, worked extensively with GPUs, and used his celebrity status to get private looks at Google and OpenAI’s video models “I’ve always been kind of into computers since I was young. Then, when film started to move from analog film to digital, I became more interested in that aspect of it. The visual-effects workflow for many years has included machine learning, so I can write pretty shitty Python scripts and stuff like that. “With convolutional neural networks, which were the precursors to what the transformer can do, which is much more computation simultaneously, you would do things like look at what’s called a tensor. That’s the numerical translation of a visual image in numbers, like the batch number, the frame number and the red, green and blue values of each pixel in each frame. It’s just that simple. That numeric is called a tensor. “You’d use a convolutional neural network to identify patterns that reveal what’s called edge detection or feature extraction, which is identifying patterns well enough to know, this is where the window ledge is, so we can more easily take the green-screen image out and replace it with something. “That was familiar to me early on because, prior to Artists Equity, I had a small visual-effects company. I’ve worked with GPUs a lot too. The visual-effects guys said, ‘Hey, you should see. There are a couple: Google and this other company, OpenAI, are doing really interesting stuff with transformers in video.’ “I’ve learned that I can actually just call up and go, ‘Hey, it’s Ben Affleck. Can I come see what you’re doing?’ Sometimes people say yes, to my astonishment.”
272
1,057
139
12,100
550,906
Learn Pi while using Pi! Pi School is a website and agent-driven course that shows you how Pi works and how to set it up. It now covers MCP, Code mode, local models, image generation, and lots of other little goodies. pi-school.ziki.boo
8
49
1
603
35,436
Jesus brutal TLDR
Mistral is back: -pushing the frontier on dollar per intelligence (1.13$ per task for a 38. mimo v2.6 pro gets 46 for 0.13$) -best open model out of us or europe (competition: nemotron 3 ultra 23, gemma 4 17, gpt-oss 12) -same score as gpt 6 luna. luna costs 0.07$ per task. 16x -#64 of 225 on artificial analysis. frontier -1 trillion parameters to land 3 points under GLM 5.3 flash (320B, 0.25$) -open weights (AA lists it as proprietary. weights oct 27) -most open license yet (large 3 apache 2.0, medium 3.5 modified mit, this one "custom mistral license") -only 8 points behind the best chinese open model and 20 behind opus 5.5 -beat glm 5.3 on deepswe, 62 to 61 (on mistrals chart. the live leaderboard has glm 5.3 at ~69) -state of the art in legal (15%) -highest cyber score of any model, 82% (opus 5.5 and astra refuse the task and score near zero) -same blog post brags it refuses more cyber prompts than any other open model -200M output tokens to run the index, median is 81M. le chonk talks -+29 points over large 3 (large 3 had 9) -sovereign ai (they resell glm 5.3 since august. until today the smartest model on mistrals own platform was chinese) -3 billion euro raised, 21 billion valuation, first model out of it: 38 -"no signs of saturation" (the rl run isnt finished, they shipped anyway) -the meme promised 30 trillion parameters. we got 1 (49B active) -europe can compete, said mensch. in abu dhabi -2nd of 5 in their own blind eval. behind opus 5. not 5.5
7
7
663
72,367
"Super Intelligence" leads to .si domains. .si domains are from Slovenia. Take a guess who is married to a president and is from Slovenia... Just take a guess.
191
662
98
24,657
1,504,086