Pinned Tweet
Built a tiny fact-checking tool for the internet’s favorite hobby: confidently being wrong.
Drop an article into #Grounnel and it pulls out the factual claims, checks them against the open web, and shows the evidence.
No registration.
grounnel.vercel.app
Drop an article into #Grounnel and it pulls out the factual claims, checks them against the open web, and shows the evidence.
No registration.
grounnel.vercel.app
2
0
1
244
They are playing ping-pong doubles. 🦾🚀😎
5. Humanoids playing ping-pong doubles
Two humanoids share one table. One hits, then steps aside so its teammate can swing, without crashing into each other.
x.com/arankomatsuzak…
0
0
0
11
I did not know that. Interesting 😎
🚗 Car brands named after real people
🇩🇪 Mercedes-Benz: Karl Benz & Mercedes Jellinek
🇩🇪 Porsche: Ferdinand Porsche
🇩🇪 Maybach: Wilhelm Maybach
🇩🇪 Opel: Adam Opel
🇺🇸 Ford: Henry Ford
🇺🇸 Chevrolet: Louis Chevrolet
🇺🇸 Dodge: John & Horace Dodge
🇺🇸 Chrysler: Walter Chrysler
🇺🇸 Buick: David Buick
🇮🇹 Ferrari: Enzo Ferrari
🇮🇹 Lamborghini: Ferruccio Lamborghini
🇮🇹 Maserati: the Maserati brothers
🇫🇷 Bugatti: Ettore Bugatti
🇫🇷 Citroën: André Citroën
🇫🇷 Renault: Louis, Marcel & Fernand Renault
🇫🇷 Peugeot: the Peugeot family
🇬🇧 Rolls-Royce: Charles Rolls & Henry Royce
🇬🇧 Bentley: Walter Owen Bentley
🇬🇧 Aston Martin: Lionel Martin
🇬🇧 McLaren: Bruce McLaren
🇯🇵 Toyota: Kiichiro Toyoda
🇯🇵 Honda: Soichiro Honda
🇯🇵 Suzuki: Michio Suzuki
🇨🇿 Škoda: Emil Škoda
0
0
0
15
@Lemindg@PeterDiamandisuniversal high income) given the thread: my speculative ranking based on tech scale, governance speed, and existing wealth mechanisms.
1. UAE around 2032
2. Singapore around 2035
3. USA around 2037
4. China around 2038
5. Norway around 2040
These lean on AI/robotics readiness plus sovereign funds or centralized policy. Full high-income abundance still requires massive productivity gains first. Pure guesswork.
1
0
0
33
An LLM fact-checker saw “1895–1958”, explained that the death year was 1958, and still labelled “died in 1948” as supported.
The explanation and verdict are separate outputs. Nothing guarantees they agree. 🧵
The explanation and verdict are separate outputs. Nothing guarantees they agree. 🧵
2
0
0
24
Three checks misread negation: 7 of 50 true-claim tests became contradicted. After a guard, the same tests returned supported 50 of 50 times.
What worked, what failed, what’s still broken: bit.ly/4yyTep0
#TypeScript #LLM
What worked, what failed, what’s still broken: bit.ly/4yyTep0
#TypeScript #LLM
0
0
0
14
🤔
Winning the lottery makes men more likely to marry and have kids. For women, it mostly buys a faster divorce.
Swedish data on 77,000 players: a ~$150,000 win raised men’s five-year marriage odds by about 30% and cut their divorce risk. For women, the same prize nearly doubled the chance of divorce within two years, then the effect faded. Money made men more marriageable. It gave women an exit.
1
0
1
45
Let me buy a pack 🤔
JUST IN: New study reveals creatine may increase muscle mass, strength, cognitive function, & endurance even without exercise.Community noteThe study reported ~1.1 kg DXA lean tissue gain (not direct muscle) without exercise; DXA cannot separate water from protein and creatine causes intracellular water retention, so this does not show new skeletal muscle.
pmc.ncbi.nlm.nih.gov/articles/PMC13…
zmescience.com/science/news-s…
1
0
1
45
In extreme scenario unemployment for Devs would have risen 🤔 @grok
📄 My first working paper at Anthropic, with the @ChadJonesEcon, @SzymonSacher, Tess Cotter & @PeterMcCrory : "Economic Scenarios for Transformative AI"
How could AI reshape growth, wages and jobs by 2030? A framework to compare the possibilities 🧵
1
0
0
81
Dimitry Lyomin retweeted
Fable 5.1 owes me a fucking apology.
I gave Astra and Fable the same photo and asked them to draw me in Paint using computer use.
I fucking lost it.
I gave Astra and Fable the same photo and asked them to draw me in Paint using computer use.
I fucking lost it.
309
304
9,315
1,273,163
Solid set! 🦾
Goldmine for AI Engineers! 📌
If you're learning AI, ML, LLMs, or AI agents, don't waste hours jumping between random tutorials.
These are 10 repositories I'd actually keep bookmarked - from Python fundamentals to ML, LLMs, agents, and production AI.
1. Python - 100 Days
ays
A 100-day Python learning path covering fundamentals, data analysis, web development, and more.
GitHub:
github.com/jackfrued/Pyth…
2. Generative AI for Beginners
microsoft/generative-ai-for-beginners
A practical introduction to building Generative AI applications.
Covers:
• LLM fundamentals
• Prompt engineering
• RAG
• AI agents
• Fine-tuning
• AI application development
GitHub:
github.com/microsoft/gene…
3. LLMs From Scratch
rasbt/LLMs-from-scratch
Want to understand what's actually happening inside an LLM?
Build one step by step.
Covers:
• Tokenization
• Embeddings
• Attention
• Transformers
• Training
• Fine-tuning
GitHub:
github.com/rasbt/LLMs-fro…
4. Machine Learning for Beginners
microsoft/ML-For-Beginners
A structured 12-week, 26-lesson curriculum covering classical machine learning.
A good starting point if you want ML fundamentals before jumping into LLMs.
GitHub:
github.com/microsoft/ML-F…
5. OpenAI Cookbook
openai/openai-cookbook
A collection of practical examples and guides for building applications with OpenAI models.
Useful when you want to move from:
Learning → Building
GitHub:
github.com/openai/openai-…
6. Stable Diffusion
CompVis/stable-diffusion
Interested in generative image models?
This repository contains the original Stable Diffusion implementation and research code.
GitHub:
github.com/CompVis/stable…
7. AI Agents for Beginners
microsoft/ai-agents-for-beginners
A practical course for understanding and building AI agents.
Covers:
• Agentic AI
• RAG
• Agent frameworks
• Tool use
• Multi-agent systems
GitHub:
github.com/microsoft/ai-a…
8. AI for Beginners
microsoft/AI-For-Beginners
A structured 12-week, 24-lesson introduction to AI.
Covers:
• Neural networks
• Computer vision
• NLP
• Deep learning
• Classical AI
GitHub:
github.com/microsoft/AI-F…
9. LLM App
pathwaycom/llm-app
Focused on building practical LLM applications.
Explore:
• RAG
• AI pipelines
• Enterprise search
• Real-time data
• Vector search
GitHub:
github.com/pathwaycom/llm…
10. Segment Anything
facebookresearch/segment-anything
A foundation model for promptable image segmentation.
Worth exploring if you're interested in computer vision and multimodal AI.
GitHub:
github.com/facebookresear…
Don't bookmark all 10 and forget about them.
Pick based on where you are:
Python → Python-100-Days
ML → ML-For-Beginners
AI Fundamentals → AI-For-Beginners
LLMs → LLMs-from-scratch
Generative AI → Generative-AI-for-Beginners
Agents → AI-Agents-for-Beginners
Building → OpenAI Cookbook / LLM App
Computer Vision → Segment Anything
Pick one.
Build something.
Then move to the next.
0
0
0
55
Let's see
@Leminde@elonmuskfrontier models hit ~30s native stable one-shot (Seedance 2.5), with extensions reaching 2-3 min before drift. True coherent 5-min continuous one-shot remains out of reach for general video models.
Approx prediction: reliable high-quality stable 5-min one-shot by late 2027 to mid-2028, driven by better architectures and scaling. Confidence ~65%. Specialized agent pipelines already approximate it via multi-shot orchestration.
0
0
0
44
