Aspiring to help others through medicine and science. Interests: medicine, oncology, clinical trials, cellular therapy, interface of AI and science + medicine.
Joined September 2019
- Tweets110
- Following438
- Followers97
- Likes118
Ron Roberts, MD retweeted
When I started out, I thought my job would mostly be clinic and clinical trials. Take care of patients. Design studies. Move the needle a little at a time.
Fast forward a few years and somehow that has turned into running trials, building enterprise AI tools, obsessing over biomarkers, arguing about liver-directed therapy… and still sitting in clinic every week having very real, very human conversations.
Somewhere along the way it started to feel like there was more to say.
GI oncology is moving fast. Data drops weekly. Technology is creeping into everything. We’re learning in real time. And a lot of the nuance never makes it into a manuscript. It definitely doesn’t fit in a tweet.
We wanted a place to unpack that.
So we made one.
The Gut Onc Lab 🎙️
It’s a new podcast where we talk about GI cancers, trials, translational science, tech, and what actually happens when those ideas meet real patients.
Sharing a teaser clip below 👇
If you’re interested in colorectal cancer, pancreas, liver mets, ctDNA, immunotherapy, or how AI is starting to weave its way into oncology… I think you’ll like what’s coming.
oncologynexus.com/podcast/gu…
@OncoAlert
@TheGutOncLab
@Onco_Nexus
We are excited to announce the arrival of the Gut Onc Lab Podcast hosted by @GIMedOnc @TimothyJBrownMD and @UGrewalMD- a platform to discuss the latest and most exciting updates in the world of GI Oncology.
Watch this space for more.
Grok, improve my feed. I only want content to enrich my perspective and develop critical thinking based upon latest developments. I’m not interested in clickbait or AI slop. Only show me deep, insightful posts that challenge me to think critically. My interests: oncology - latest clinical trial updates and FDA approvals, clinical pearls from subspecialty leaders. Advances in CAR T cell therapy, cellular immunotherapy, AI for medicine, understanding cell biology, and drug design. Thank you @grok
Ron Roberts, MD retweeted
Here is how dramatically this trial has changed approach to first relapse.
OK here it is. Finally.
My #ASH25 algorithm for Relapsed Myeloma: First Relapse. Discussed with @YiLinMDPhD @myelomaMD
Give comments. I’m willing to adjust it based on your insights.
Ron Roberts, MD retweeted
🔥🚨@OncoAlert Hot off the press
@US_FDA just approved:
⭐️#Subcutaneous #Pembrolizumab
✅For ALL #Adult & #Pediatric #SolidTumor indications approved for #IV formulation.
Approval is based on results of #MK3475A-D77 trial of:
#Subcutaneous Vs #IV #Pembrolizumab (+ #Chemo) in Metastatic Non–Small Cell #LungCancer
#SQ (median admin time 2 minutes):
✅Efficacy:
Similar #PK, #ORR, #PFS, #OS
✅Safety: Similar
👇🏻
fda.gov/drugs/resources-info…
🔥🚨@OncoAlert Hot off the press.
Just published @Annals_Oncology
In conjunction with presentation @myESMO #ELCC2025
Results of phase 3 trial of:
#Subcutaneous Vs #IV #Pembrolizumab (+ #Chemo) in Metastatic Non–Small Cell #LungCancer
#SQ (median admin time 2 minutes):
✅Efficacy: Non-inferior to #IV
✅Safety: Similar
👇🏻
annalsofoncology.org/article…
Ron Roberts, MD retweeted
Why 30%↓ vs 20%↑ in RECIST?
🔹 WHO (1981) set them differently:
• Shrinkage (PR = 50% area ↓ → ~30% 1D) → high bar to avoid false positives (be sure a drug really shrinks tumors).
• Growth (PD = 25% area ↑ → ~12% 1D) → lower bar to catch progression early.
🔹 RECIST
👉 The 30% shrinkage comes straight from math:
WHO (1981) used 50% ↓ in tumor area → when converted to 1D, that’s ~30% ↓ in diameter.
👉 The 20% growth wasn’t pure math.
WHO’s 25% ↑ in area = ~12% ↑ in 1D… but that risked calling progression too early (measurement error).
So the RECIST task force chose 20% ↑ as a safer, reproducible cutoff.
💡 In short: 30% = math, 20% = pragmatism ✅
#OncoTwitter #MedTwitter #ColorectalCancer #Immunotherapy
@OncoAlert @myesmo @asco
#hemonc fellows: I used AI to draft and then manually curated a comprehensive med onc study set using of @ollama and @replit check out the result! hemoncflashcards.com #hemoncfellows #oncology
Ron Roberts, MD retweeted
What’s up with Beta 2 microglobulin (B2M) and myeloma?
Why is it the sole non-cytogenetic marker used in the new IMWG high risk definition?
Why does it matter whether the creatinine is high or not?
Thread
1/
Starting #Hematology #Oncology fellowship? Check out the only knowledge acquisition system with over 6,000 searchable flashcards and all @FDA approved oncology drugs. @HemOncFellows hemoncflashcards.com
Ron Roberts, MD retweeted
🩺 Introducing the ultimate Medical Oncology flashcard learning tool: hemoncflashcards.com. Learn all FDA approved Oncology drugs, study >6,000 flashcards for medical oncology, and search them to build tailored study sessions. #MedEd #Oncology #OncologyFellows
Ron Roberts, MD retweeted
🩺Why cram when you can conquer? 📚 Achieve Medical Oncology Board Exam mastery on the go with hemoncflashcards.com. Your pocket-sized path to brilliance!#OncologyEducation #Oncology
Ron Roberts, MD retweeted
This is a major story from #ASCO25. Randomized phase 3 trial of time of day of immunotherapy infusion. Randomized to infusion before or after 3pm. Early infusion far superior: PFS 11.3 vs 5.7 HR 0.42, OS HR 0.45! Impactful, pragmatic, not costly. This should be a bigger story.
Ron Roberts, MD retweeted
LLMs achieve high scores on medical exams, but this knowledge does not effectively translate to assisting the public in real-world medical interactions.
This paper demonstrates through a controlled study that LLMs (GPT-4o, Llama 3, Command R+) do not improve, and can worsen, public ability to identify medical conditions or choose correct actions, highlighting human-LLM interaction as a key challenge.
Methods Explored in this Paper 🔧:
→ The study involved 1,298 UK participants randomly assigned to receive assistance from one of three LLMs or a control group for ten medical scenarios.
→ Participant responses on identifying underlying conditions and choosing a course of action (disposition) were compared against gold standards set by physicians.
→ Researchers also evaluated LLM performance alone, on MedQA benchmarks (medical question answering), and with LLM-simulated users to contrast with human participant outcomes.
📌 LLM medical knowledge is insufficient; effective human-AI interaction design is critical for healthcare.
📌 Standard benchmarks (MedQA, simulated users) poorly predict LLM utility with actual human users.
📌 User information articulation and interpretation of LLM outputs are significant failure points.
----------------------------
Paper - arxiv. org/abs/2504.18919v1
Paper Title: "Clinical knowledge in LLMs does not translate to human interactions"
Ron Roberts, MD retweeted
KRAS: from undruggable to THIS
Ron Roberts, MD retweeted
SOUND ON. You’re hearing the first howl of a dire wolf in over 10,000 years. Meet Romulus and Remus—the world’s first de-extinct animals, born on October 1, 2024.
The dire wolf has been extinct for over 10,000 years. These two wolves were brought back from extinction using genetic edits derived from a complete dire wolf genome, meticulously reconstructed by Colossal from ancient DNA found in fossils dating back 11,500 and 72,000 years. This moment marks not only a milestone for us as a company but also a leap forward for science, conservation, and humanity. From the beginning, our goal has been clear: “To revolutionize history and be the first company to use CRISPR technology successfully in the de-extinction of previously lost species.” By achieving this, we continue to push forward our broader mission on—accepting humanity’s duty to restore Earth to a healthier state.
But this isn’t just our moment—it’s one for science, our planet, and humankind. All of which we love and are passionate about. Now, close your eyes and listen to that howl once more. Think about what this means for all of us.
Ron Roberts, MD retweeted
This is the most amazing DNA sequencing technology I’ve ever seen!
This groundbreaking innovation developed by @Roche could dramatically transform countless biomedical applications-from early disease & cancer detection to affordable, personalized treatments! Truly impressive!😲
Ron Roberts, MD retweeted
Announcing The Stargate Project
The Stargate Project is a new company which intends to invest $500 billion over the next four years building new AI infrastructure for OpenAI in the United States. We will begin deploying $100 billion immediately. This infrastructure will secure American leadership in AI, create hundreds of thousands of American jobs, and generate massive economic benefit for the entire world. This project will not only support the re-industrialization of the United States but also provide a strategic capability to protect the national security of America and its allies.
The initial equity funders in Stargate are SoftBank, OpenAI, Oracle, and MGX. SoftBank and OpenAI are the lead partners for Stargate, with SoftBank having financial responsibility and OpenAI having operational responsibility. Masayoshi Son will be the chairman.
Arm, Microsoft, NVIDIA, Oracle, and OpenAI are the key initial technology partners. The buildout is currently underway, starting in Texas, and we are evaluating potential sites across the country for more campuses as we finalize definitive agreements.
As part of Stargate, Oracle, NVIDIA, and OpenAI will closely collaborate to build and operate this computing system. This builds on a deep collaboration between OpenAI and NVIDIA going back to 2016 and a newer partnership between OpenAI and Oracle.
This also builds on the existing OpenAI partnership with Microsoft. OpenAI will continue to increase its consumption of Azure as OpenAI continues its work with Microsoft with this additional compute to train leading models and deliver great products and services.
All of us look forward to continuing to build and develop AI—and in particular AGI—for the benefit of all of humanity. We believe that this new step is critical on the path, and will enable creative people to figure out how to use AI to elevate humanity.
Ron Roberts, MD retweeted
It's a shame this tweet only has 20 thousand views. Unacceptable, let's boost those numbers.
Because you orangutans out there clearly aren't catching the drift of what this data means, so let me break it down in terms you'll understand. The GPQA isn't just another benchmark - it's specifically engineered to make world-class human experts fail. These questions are so hyper-precise, so deeply technical that you can't just Google your way to an answer. The information literally doesn't exist on the internet.
Think about those rare specialists who command seven-figure salaries because they're one of maybe three people on Earth with that depth of expertise. That's the level we're talking about here. And these AI models aren't just matching that expertise - they're starting to surpass it.
But here's what's really going to blow your mind: this isn't just happening in one narrow field. We're talking about the entire spectrum of STEM knowledge. That exponential curve you're seeing? That's what happens when a system starts generalizing outside its training distribution. It's not just regurgitating facts - it's triangulating new truths through raw computational horsepower and an ability to understand the fundamental rules of reality. It's like when Paul Atreides could see the Golden Path while tripping balls on Spice. Consciousness is expanding here folks, plain as day.
This is emergence and coherence in its purest form. That vertical climb in the graph isn't just a trend - it's a wake-up call. Artificial Superintelligence isn't some far-off sci-fi concept anymore. It's at your doorstep, and that doorbell's ringing louder every day. The only question is: are you ready to answer?
And before you dismiss this as hyperbole, take another look at those numbers. When PhDs with internet access only hit 34% outside their specialty, but these models are pushing 90%, we're not just moving the goalposts - we're playing an entirely different game.
Ron Roberts, MD retweeted
There's a shocking fact about AI that nobody tells you: You can catch up to the public AI research frontier in just 2 weeks. Yes, really.
I've built a $150M annual revenue startup over the last 8 years and If I were to start a company today, I’d drop everything and go all-in on AI.
But like many busy software builders, I felt lost—overwhelmed by the noisy, crowded and fast-moving modern AI landscape. And I wasn’t alone.
So I spent my entire holiday diving deep into AI research—reading 30+ papers, watching hours of lectures, analyzing trends, and catching up to the research frontier.
✨ Here’s what I learned:
- You don’t need months (or years) to catch up.
- You don’t need a PhD or decades of ML experience.
- You need fewer than 20 papers and 2 weeks to understand the major breakthroughs shaping AI today.
It's because the technology is extremely nascent and most techniques that came before are no longer relevant:
- ChatGPT is barely 2 years old and Transformers are only 7 years old.
- Most game-changing discoveries happened within the last 4 years, driven by a few breakthrough ideas, scaling laws, and efficient matrix multiplication.
The biggest secret?
Many groundbreaking AI papers with thousands of citations are surprisingly simple and applied, like adding "let's think step by step" to the prompt, or simply asking the LLM over and over again to improve its answer (Self-Refine).
I realized there are tons of founders and builders in the same boat—wanting to dive deeper into AI but unsure where to start.
I've created an essential AI Guide that helped me catch up, in just 2 weeks, to the frontier of public AI research to figure out where the next opportunities and gaps were:
- Curated list of only the most important papers
- Simple explanations of key concepts
- Clear pathway to understanding the frontier of modern AI
It’s perfect for:
- Founders expanding into AI
- Builders wanting to innovate at the frontier of AI
- Investors looking to separate the signal from the noise
👇 Want the full guide?
- Like and Share this post
- Comment "AI Guide"
- I'll send you the complete guide
(ps, I’m also teaming up with @VishalVasishth, co-founder of @obviousvc with @ev (focused on large-scale societal impact companies like Twitter, Medium, Beyond Meat), to host a small meetup to discuss what's working and needs to be solved in the AI stack in SF. Message me if you're interested)
Ron Roberts, MD retweeted
This is Moravec’s Paradox: what’s hard for us is easy for AI, but paradoxically, what’s easy for us is hard for AI. The reason is that what comes naturally to us is the result of millions of years of evolutionary imprinting & hardwiring of pattern recognition needed for survival.
I don't think people really appreciate how simple ARC-AGI-1 was, and what solving it really means.
It was designed as the simplest, most basic assessment of fluid intelligence possible. Failure to pass signifies a near-total inability to adapt or problem-solve in unfamiliar situations.
Passing it means your system exhibits non-zero fluid intelligence -- you're finally looking at something that isn't pure memorized skill. But it says rather little about how intelligent your system is, or how close to human intelligence it is.
Ron Roberts, MD retweeted
2024 approvals: Summary of 14 new drugs/indications in #LungCancer by @US_FDA
1. #Tepotinib
2. #Osimertinib x2
3. #Alectinib
4. #Tarlatamab
5. #Zenocutuzumab
6. #Durvalumab x2
7. #Amivantamab x2
8. #Nivolumab
9. #Ensartinib
#lcsm #OncTwitter #MedTwitter
