Eka’s robotic claw feels like we're approaching a ChatGPT moment
OpenAI’s o1 correctly diagnosed 67% of ER patients vs. 50-55% by triage doctors
By Saeed Ezzati7 min read
In today’s email:
👨🏻⚖️ Inside Musk Vs OpenAI battle
🫧 Maybe AI Isn't a Bubble After All
🧠 Will human minds still be special in an age of AI?
🧰 10 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.
Eka’s robotic claw feels like we're approaching a ChatGPT moment
Key Takeaway: Eka’s touch-sensitive robotic claw suggests that robots may be approaching a breakthrough in real-world dexterity, using simulation-based learning to handle delicate, unpredictable tasks once thought nearly impossible.
More Insights:
Eka’s robot can screw in a light bulb, pick up keys, handle hairbrushes, and sort chicken nuggets with unusually natural movement.
The startup’s founders believe dexterity can be solved by scaling simulation-based training, rather than relying mainly on human demonstration videos.
Eka uses custom grippers with touch sensing and a “vision-force-action” model that accounts for physics like weight, speed, mass, and inertia.
The article compares Eka’s robots to early GPT models: imperfect, but showing signs of a much bigger intelligence breakthrough.
If successful, this technology could transform factories, food service, warehouses, restaurants, and eventually homes.
Why it matters: Eka’s work matters because the next AI revolution may not be about what machines can say, but what they can physically do—and once robots master the human hand, they may begin reshaping labor, industry, and everyday life in ways language models never could.
One brand built 30+ landing pages through Viktor without a single developer.
Each page mapped to a specific ad group. All deployed within hours. Viktor wrote the code and shipped every one from a Slack message.
That same team has Viktor monitoring ad accounts across the portfolio and posting performance briefs before the day starts. One colleague. Always on. Across every account.
5,700+ teams. 3,000+ integrations.
OpenAI’s o1 correctly diagnosed 67% of ER patients vs. 50-55% by triage doctors
Key Takeaway: A Harvard-led study found OpenAI’s o1 model outperformed two internal medicine doctors in diagnosing real ER triage cases, but researchers warn AI is not ready to make life-or-death decisions without human oversight.
More Insights:
OpenAI’s o1 gave an exact or very close diagnosis in 67% of triage cases, compared with 50–55% for the two physicians.
The AI performed best when information was limited and urgency was highest, such as initial ER triage.
Researchers stressed the study only tested text-based medical records, not visual cues, patient distress, or bedside judgment.
The study calls for real-world clinical trials before AI is used broadly in patient care.
Critics noted the comparison was against internal medicine physicians, not ER specialists, which may make some headlines overstated.
Why it matters: AI may become a powerful second-opinion tool in medicine, but the real challenge is not whether it can guess diagnoses—it is whether healthcare systems can use it responsibly without weakening human accountability, judgment, and patient trust.
Inside Musk Vs OpenAI battle
Key Takeaway: Elon Musk’s courtroom battle with OpenAI is exposing allegations that he wanted control of OpenAI, used insiders to track it after leaving, and may have relied on OpenAI’s own models to help build rival startup xAI.
More Insights:
Musk appeared to admit xAI “partly” used OpenAI models through distillation, while arguing that AI labs commonly use competitors’ systems to train or validate their own.
OpenAI’s lawyers are framing Musk’s lawsuit as less about AI safety and more about weakening a rival after he failed to control OpenAI.
A new OpenAI filing claims Musk sent threatening texts to Greg Brockman and Sam Altman shortly before trial, warning they would become “the most hated men in America.”
Court evidence portrays Shivon Zilis as a key behind-the-scenes bridge between Musk and OpenAI, sharing updates and managing relationships even after Musk left OpenAI’s board.
The judge appeared skeptical of Musk’s timing, especially because his concerns intensified around the same period he launched xAI as a for-profit competitor.
Why it matters: This trial is not just about OpenAI’s corporate structure; it is becoming a public autopsy of how the AI race is really fought—through money, control, talent poaching, insider access, model imitation, and competing claims of “safety” that may also serve business interests.
More:
- A profile of William Savitt, Sam Altman's lead lawyer against Elon Musk
Other stuff
A Dark-Money Campaign Is Paying Influencers to Frame Chinese AI as a Threat
Maybe AI Isn't a Bubble After All
Will human minds still be special in an age of AI?
A profile of OpenAI CFO Sarah Friar, who sources say helped keep OpenAI's Microsoft deal on track
AI discovery reveals DNA isn’t locked away in cells after all
AI Has Made Memory Chips One of the World’s Most Profitable Products
Take ChatGPT to the next level 🎉
Add folders and subfolders, prompt manager, prompt optimizer, image gallery, side-by-side voice mod, PDF export, reference chats, chat notes, and many more features.
Get the extension for FREEEditor's Pick
Codex Pets - Animated companions for your Codex workflow
Marx Finance - AI agents debate the markets
The best prompt engineers aren't typing. They're talking.
Power users figured this out early: speaking a prompt gives you 10x more context in half the time. You include the edge cases, the examples, the tone you want — because talking is fast enough that you don't skip them.
Wispr Flow captures everything you say and turns it into clean, structured text for any AI tool. Speak messy. Get polished input. Paste into ChatGPT, Claude, Cursor, or wherever you work.
89% of messages sent with zero edits. 4x faster than typing. Works system-wide on Mac, Windows, and iPhone.
Buda - Recruit agents to run your company as a synchronous team
nudge - Drop your tasks. AI auto-schedules your whole week.
Montage - The runtime framework for agentic user interfaces!
Scholé - Turn everyday work into personalized AI learning
Feather - Photo editor with local AI
Filect - Organize Your Files With AI
PandaProbe - open source agent engineering platform
Rosentic - Catch when coding agents break each other before merge
Unclassified 🌀
WFH Team - Work from anywhere in the world
