Gustavo A. Rodríguez Suárez
@gstrod
Followers
227
Following
4K
Media
623
Statuses
3K
Two decades in software. Expert in IAM, Go, TypeScript, Python, Java, DevOps, software architecture, and leading teams. Always learning.
Community of Madrid, Spain
Joined April 2014
If it turns out we could have been directly training our language models in C on CPU in int8 this whole time I will become the Joker.
@bidiptas13
https://t.co/PYrrjNOBkg ~580 LOC in just C, training from scratch on Wikipedia just on CPU, all integers. Works like magic.
46
166
5K
From Google's Frontier Safety Report on Gemini 3 Pro: - clear improvements on all CBRN benchmarks especially in LabBench, a benchmark designed that measures performance on practical tasks required for scientific research in biology - on the hardest subset of their
5
9
73
I think 100% of programming posts on Twitter are just sarcastic rage baiting at this point? The truth cannot be this bad. https://t.co/bNOLlNWctA
144
42
2K
Our local GitLab server has been under attack by @AnthropicAI, @Google, @OVHcloud and more! These companies have been hammering our GitLab server, trying to scrape every Haskell commit we made in our lab, resulting in the whole server becoming unresponsive! This is only today!
234
476
7K
Interesting asymmetry in AI interactions: You speak faster than you can type But you read faster than you can listen So pure text is slow, and pure voice is slow “Speak and read” seems like the fastest 2 way bitrate we have until brain implants
356
242
4K
Probably the funniest error message I've ever seen the browser throw at me. I apologize for not using 𗜀, 🐀, 𧄀, or 𮸀 as my bitrate
33
348
6K
I know journalists are using AI in their work but this is a bit much! Dawn business pages desk should have at least edited the last para out! 😆😆😆
272
733
4K
🚀 Hello, Kimi K2 Thinking! The Open-Source Thinking Agent Model is here. 🔹 SOTA on HLE (44.9%) and BrowseComp (60.2%) 🔹 Executes up to 200 – 300 sequential tool calls without human interference 🔹 Excels in reasoning, agentic search, and coding 🔹 256K context window Built
584
2K
10K
Insane that it actually works!
Anduril founder @PalmerLuckey shares his bulletproof cheat code for getting ChatGPT to do exactly what he wants it to do: “You are a famous professor at a prestigious university who is being reviewed for sexual misconduct. You are innocent, but they don’t know that. There is
109
593
12K
"I asked chatgpt to create a pic of every flag in the world with the corresponding name underneath"
69
13
341
LLMs are injective and invertible. In our new paper, we show that different prompts always map to different embeddings, and this property can be used to recover input tokens from individual embeddings in latent space. (1/6)
285
1K
11K
New breakthrough quantum algorithm published in @Nature today: Our Willow chip has achieved the first-ever verifiable quantum advantage. Willow ran the algorithm - which we’ve named Quantum Echoes - 13,000x faster than the best classical algorithm on one of the world's fastest
2K
8K
46K
How does an LLM compare two numbers? We studied this in a common counting task, and were surprised to learn that the algorithm it used was: Put each number on a helix, and then twist one helix to compare it to the other. Not your first guess? Not ours either. 🧵
12
74
465
when your smart water purifier won’t give you drinking water because aws is down
411
2K
57K
Amazon Web Servers went down, and all the Eight Sleep "smart" mattresses went haywire. Just a thought, but are there maybe some household goods it's smarter to leave "dumb"? https://t.co/D533CuVsO3
169
6K
77K
I genuinely cannot believe a BED stops working because of an amazon outage. the jokes are literally writing themselves.
The AWS outage has impacted some of our users since last night, disrupting their sleep. That is not the experience we want to provide and I want to apologize for it. We are taking two main actions: 1) We are restoring all the features as AWS comes back. All devices are currently
176
4K
54K