Samar Abedrabbo, PhD

@drkeithsiau Enjoy your time in Egypt Dr. Siau, maybe by some crazy coincidence will run into you, will be there throughout next week with my family 🙂

Grateful to be invited as faculty at the WEGE Summer School in Cairo. Huge credit to Nada El-Domiaty and Shimaa Afify for leading this pioneering initiative to empower female endoscopists and elevate quality of endoscopy in Egypt. Thank you for the warm welcome! 🙏🙌

@andrewwhite01 @Nature @agreeb66 Well-deserved, congrats Andrew and team 🙂

Our agents can make real discoveries! One year of peer-review, but this work was accepted in @Nature . Huge congrats to the team, especially @agreeb66 who patiently worked on follow-up experiments for months. And here's a pics of some authors in the lab where we did the work

As long as it’s not a self verification loop 🙈 when a model is grading its own work, it mostly produces the same answer back like patting itself on the back… the blind spot that caused the first error are usually still there on the second pass, which is why when it double checks itself it doesn’t really catch anything significant… either use a structurally different agent like a different model or a different role or ask it to run actual tools to verify 🙂

@samar_abedrabbo verification loops make AI actually useful instead of just flashy... what's been working best for you?

Wrote my first Substack: I put Claude Code, Codex CLI, and Grok Build in one terminal and ran them across 5 real biology tasks and individually. The piece is about how AI agents fail at biology/biomed tasks, what actually fixes those failures, and why verification loops matter more than just adding more tools. Excited to get back to writing my own thoughts and runs on AI evaluations/benchmarks for biology, medicine, STEM, and education more broadly :) "The Biology Terminal: What Happens When Frontier AI Agents Are Asked to Do Real Science" open.substack.com/pub/samarabedr…

@DeryaTR_ @OpenAI @sama 🤍

Met a really cool person at an AI party😊 @openai @sama

Omg I love this question :) the best way I would describe it is it’s like a little bug that goes inside your body and starts using your own nutrients and your own body’s ways of doing things to make more copies of itself, but this is not good because it also causes your body to do a lot of inflammation… some viruses are worse than others, depending on the types of body cells they target in you

Today, a very basic question for the med community: If you had to explain to a patient what a VIRUS actually is - without using any medical jargon - what is the one analogy you’d use?

@sama Make research/discovery/tool-using models smarter/more expensive and regular models for everyday questions cheaper/faster…🙃

i keep thinking i want the models to be cheaper/faster more than i want them to be smarter but it seems that just being smarter is still the most important thing

@nattyover But the fun question is, is it moving clockwise or counterclockwise and what is the significance of each? 😉

Bacteria move around using a molecular machine called the flagellar motor that rotates faster than the flywheel of a race car engine and switches directions in an instant. After 50 yrs, scientists have finally figured out how it works. “My lifelong quest is now fulfilled.” Link⤵️

Our life sciences team tagline: "Scale test-time compute to cure all disease"

Found my favorite childhood book growing up 🥰I remember when I discovered this gem when I was around 10 years old in a little library💜

🥳🧬🧫🥰AI 🚀

To go deeper on our new Life Sciences model series, research lead @joyjiao12 and product lead Yunyun Wang joined @AndrewMayne on the OpenAI Podcast to discuss how we’re building models for biology, drug discovery, and translational medicine. They cover both the opportunity and the responsibility ahead: better research workflows today, more autonomous labs over time, and careful deployment from day one.

@kevinweil Congratulations guys 🥳 🧬🫶🏻

See our blog post for videos and demos: openai.com/index/introduc… Request access to GPT-Rosalind here: openai.com/form/life-scie… Life Sciences plugin: github.com/openai/plugins…

💥 Super excited to launch GPT-Rosalind, our first frontier model built for scientific research across biology, drug discovery, and translational medicine. The model is trained in chemistry, protein engineering, genomics and more, including built-in knowledge of relevant databases and tools researchers use. We want to make these capabilities available to scientists and researchers while maintaining strong safeguards against biological misuse. GPT-Rosalind is launching through a trusted access deployment structure for qualified customers. We're also launching a Life Sciences plugin for Codex to everyone today, which works with our mainline models as well as GPT-Rosalind.

Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.

@cgtwts 🤍🤍

Jensen Huang, CEO NVIDIA: "Claude is incredible. Anthropic has made huge progress, a massive leap, in developing Claude. We use it all over our company. The coding capability, the reasoning capability, its overall ability is genuinely impressive” “On the other hand, ChatGPT is probably the most successful consumer AI in history. Its ease of use and approachability mean everybody should get involved” how often do you see a man running $NVDA openly talk about what he missed rare, and that’s exactly what separates the great ones.

@HarryStebbings @AnjneyMidha Anthropic has great culture, it’s well known in the Bay Area and cares about their people, they’re also brilliant in stem because they value science.

Dario Amodei is arguably the second-greatest entrepreneur of our time (thanks, Elon). So what makes Dario so special? @AnjneyMidha, a man who knows him as very few do, reveals the three things that make Dario the $TRN founder. 👇 1. Scientific brilliance. 2. He has an obsessive desire to run empirical experiments. 3. Mission alignment and culture. No drift. Focus. That attracts the best talent.

@tszzl Use Claude

tried to use clankers for my taxes and they mostly made it harder not easier. we’ve got a ways to go

@OpenAINewsroom 🧬🙌

~200,000 people each week are using ChatGPT for advanced life sciences work, according to data studied through a privacy-preserving automated pipeline. More insights and use cases in our new report: cdn.openai.com/pdf/93f630b9-e…

🚀 Today we're announcing Amazon Bio Discovery, a new AI-powered application designed to help scientists design and test novel drugs more quickly and confidently. Scientists get access to AI models for drug discovery, and can converse naturally with an AI agent to select the right models for their research needs, optimize inputs, and evaluate drug candidates. They then send the top candidates to physical labs for testing, with results routing back for rapid iteration. Here's how we're making AI accessible to scientists in the lab ➡️ aboutamazon.com/news/aws/aws-a…

@DrNeilStone But you will always remember how much you loved the Kreb’s cycle and how you use it daily in your career 😉

You will forget most of what you learned at med school and start learning on the job once you actually qualify