Causal AI Research Scientist @ToyotaResearch | CS PhD @UCLA
Mountain View, CA USA
Joined December 2007
- Tweets325
- Following245
- Followers525
- Likes16.5K
The first edition of this book sparked my love of causal inference and my infinite admiration and respect for Judea Pearl. Soon after, I had the extraordinary privilege of becoming his PhD student.
Scott Mueller retweeted
Happy Publication Day!
According to my publisher, the 2nd edition of The Book of Why should now be on the shelves.
Can't wait to get my own hard copy -- it's always a soft touch of "love at second sight."
Would you bet your startup on AWS?
@awscloud shut off our Bedrock & Lambda over a false compromise alert. We changed our root password and checked CloudTrail as asked.
2 days of silence. 10 follow-ups. Waited hours for a chat agent, then gave up. Repeatedly. @AWSSupport
It's absurd that AWS hasn't given even a basic response to a critical support ticket after TWO WHOLE DAYS.
@awscloud, you sell AI agents. Can't staff support? Use one. It could at least acknowledge the ticket and tell us what happens next. @AWSSupport
For the record, @AWSSupport replied to this within 13 minutes, just after midnight. 3 days in, AWS still hasn't fixed its own mistake. xitter.cf/smueller/status/210791…
AWS has answered my post, my DMs, and my ticket. It still hasn't fixed its own mistake.
It wrongly flagged our account as compromised and shut off our Bedrock & Lambda. 3 days later, we're still down.
@awscloud, maybe have AI run AWS for non-enterprise customers? @AWSSupport
AWS has answered my post, my DMs, and my ticket. It still hasn't fixed its own mistake.
It wrongly flagged our account as compromised and shut off our Bedrock & Lambda. 3 days later, we're still down.
@awscloud, maybe have AI run AWS for non-enterprise customers? @AWSSupport
I apologize for the delay in case assistance. You can share your support case via private message so that I can check that it's with the correct team, and share your sentiment internally. If you're prompted to enter our X number, there's no need; you can send us a message and reach our inbox without it. ^RZ
🙌🏼 I’m hiring a PhD research intern at @ToyotaResearch to test and improve causal reasoning in LLMs.
Bring your causal inference and LLM experience.
Winter 2027 · Paid · 12 weeks · Hybrid, Los Altos, CA
Apply or share:
jobs.lever.co/tri/6433edb6-7…
Excited to be a co-investigator in TRI's Human-AI Kaizen Initiative: tri.global/news/elevating-hu…
Against Academic Boycots - Sign the Petition! c.org/j5ZrqfxJDz via @Change
Scott Mueller retweeted
Sup AI is live on @ProductHunt 🚀
"Which AI model is the best?"
Wrong question.
The best model isn't a model. It's an orchestra.
Sup AI runs 9 frontier models in parallel and synthesizes their answers→ 52.15% on HLE benchmark (without the help of tools).
→ Multi-model consensus (up to 9 models)
→ Ensemble RAG with live web + your files
→ Every claim cited
$10 free credit to start
20% off with code: PRODUCTHUNT
Links below 👇
Sup AI just launched on @ProductHunt
Sup combines multiple AI models and uses confidence scoring to give better answers with fewer hallucinations.
#1 on Humanity's Last Exam: 52.15%. Beating every individual model.
$10 starter credit to try it
producthunt.com/products/sup…
Did @OpenAI just silently kill top-20 logprobs for all models except GPT-5.4-nano? Setting `top_logprobs` to anything > 1 returns an error for the past 4-5 hours. Might just be their servers having trouble, but their status page indicates "No incidents" for their responses API.
Scott Mueller retweeted
Sup AI whitepaper is live on the methodology behind 52.15% on HLE:
• 3 correct answers synthesized when EVERY model failed
• Grok 4 (29%) uniquely solved 16 Qs vs GPT-5 Pro's 9 (40%)
• Low correlation pairs >high accuracy pairs
• 58.44% theoretical ceiling w/ models
• 42% Qs unsolved by ANY model
• Full methodology, IQ curves, correlation matrices: sup.ai/research/hle-white-pa…
#AI #MachineLearning #OpenSource #AIResearch #EnsembleAI #AIOrchestration #HLE
Scott Mueller retweeted
Sup AI's 52.15% HLE (+7.41 over frontiers) was orchestration + synthesis.
Now every model executes Python/Bash/C++/JS/TS/R/Java +15 langs. Image mutation. Virtual FS. Deterministic verification.
Guesses → Calculations. Ceiling exploded.
#SupAI #AI #CodeExecution
Scott Mueller retweeted
That's ~$950/month across 5 services. Sup AI is $200/month and includes all those models and more in one place. Save $750/month. sup.ai
Scott Mueller retweeted
🗂️ Deprecated models are now accessible: Claude Opus 4.1, Gemini 2.5, Flash Gemini 2.5 Pro, Llama 3.3 70B, Llama 4 Maverick 17B, Llama 4 Scout 17B, Kimi K2 Turbo, Grok 4 Fast, Grok 4 Fast Reasoning, GPT-5,
GPT-5 Pro, GPT-5.1, GLM 4.5 Air, GLM 4.6, MiniMax M2, Pixtral 12B are back by request. Find them at the bottom of the model selector → click "Deprecated" to expand. Great for: specific personalities, fewer guardrails ⚠️ Not recommended for serious work as newer models outperform them.
Scott Mueller retweeted
We just launched the Sup AI Developer API
One endpoint → Multiple frontier models → Better answers
✅ Multi-Model Consensus: Combine outputs from Claude, GPT-5, Gemini, and more
✅ OpenAI compatible (2-line integration)
✅ 5 modes: fast → thinking → pro
✅ 52.15% on Humanity's Last Exam (SOTA)
✅ Self-healing tool calls
Get your API key → sup.ai/api
Full docs → docs.sup.ai
How it works:
Instead of betting on one model, Sup AI orchestrates multiple models and synthesizes their outputs.
auto mode picks the right approach. pro mode runs 9 models for mission-critical work.
You get consensus-driven answers without the infra headache.
Scott Mueller retweeted
Sup AI update: → Faster generation → More reliable → Terminate models mid-response (for when you can't wait for GPT-5.2 Pro to finish 🙂)
Also added GLM 4.7 and MiniMax M2.1 42 models. One interface. sup.ai
ALT Image or Sup AI interface with window allowing user to stop models