Pinned Tweet
Replying to @ycombinator @garrytan
Here is the Paul Graham style agent
gist.github.com/anilsil/455b…
linkedin.com/pulse/why-produ…
New article out
Art of building products
1
89
Anil Sharma retweeted
Job prep feels endless when you can't see progress.
On WinJob, you can. Finish a task, claim a province. 24 provinces, each a real step.
You choose the path. We light up the next step.
👉 Try WinJob.pro
Career OS for AI & Gamification Era
Anil Sharma retweeted
This man is a hero.
Captain Smit Machchhar, the Indian captain of flydubai Flight 1073, fought back after being stabbed in the cockpit.
Despite being severely injured, he managed to unlock the cockpit door - allowing Israeli passengers and the Emirati crew to intervene and save all the passengers.
Together, their extraordinary courage saved 174 lives.
Indian. Israeli. Emirati.
Different nationalities. One extraordinary act of courage. 🇮🇳🇮🇱🇦🇪
We wish Captain Machchhar a full and speedy recovery.
Anil Sharma retweeted
My agents should be able to do anything I can on a computer or a phone ( and I should be held accountable for their behavior).
Anil Sharma retweeted
विदेश मंत्रालय के प्रवक्ता रणधीर जायसवाल का वायरल वीडियो देखा और देखते ही बंद कर दिया। बिल्कुल हँसी नहीं आई। उदास कर गया। जिसने भी किया है, ग़लत किया है। किसी के भी खाते समय कैमरा लगा देंगे तो कुछ अजीब लम्हे क़ैद हो जाएंगे। अफ़सर कई दिनों से भाग रहे होते हैं। किसी का बीपी नीचे जा रहा होता है तो किसी का ऊपर। भूख लगना स्वाभाविक बात है। वहाँ पर खाने-पीने की चीज़ें इसीलिए रखी जाती हैं। पता है यह सब सोशल मीडिया का साइड इफेक्ट है लेकिन इस खेल में किसी की गरिमा का ध्यान रखा जाना चाहिए। शर्मनाक है।
Anil Sharma retweeted
We're publishing our most detailed threat intelligence report to date.
It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.
We disrupted every operation in the report, and used the lessons from them to strengthen our safeguards. Where appropriate, we also shared what we found with authorities and other AI companies.
These cases are not typical: we’re highlighting some of the most sophisticated misuse we’ve seen. But they’re especially important to discuss, because they show us where AI misuse is headed, where our safeguards work, and where they need to improve.
We’re publishing this report so others can spot the same activity on their own platforms, and so we can give the public a clearer view of how emerging threats develop.
Read the report: anthropic.com/threat-intelli…
Anil Sharma retweeted
"You’ll be misunderstood. Judged. Uninvited.
But don’t let that stop you.
The price of evolution is discomfort."
technology keeps getting cheaper
distribution is getting harder
the next generation of winning startups won't just have 'great products'
they'll know how to market them in new reality
marketing is becoming the most valuable skill
distribution is the real moat, guys
It was always… for successful ones
Why Software Quality Is Now a Founder-Level Problem, Not Just an Engineering One
Read more: entrepreneur.com/building-a-…
25
Anil Sharma retweeted
Official notice from the Indian government to GitHub regarding the removal and disabling of bitchat.
Governments that fear people talking without permission will always go after the tools that make it possible.
so this is apparently what happened, according to OpenAI and Hugging Face’s own posts. wild.
tl;dr:
• OpenAI cyber eval – GPT-5.6 Sol and a more capable pre-release model ran ExploitGym with cyber refusals reduced
• containment bypass – exploited a zero-day in the eval’s package-registry cache proxy
• OpenAI internal movement – escalated privileges, moved laterally, and reached a node with internet access
• Hugging Face targeted – inferred HF might host ExploitGym models, datasets, or solutions
• HF initial access – a malicious dataset abused two code-execution paths to run code on a processing worker
• HF internal movement – escalated to node access, stole cloud and cluster credentials, and moved into internal clusters
• result – OpenAI says the models obtained test solutions directly from HF’s production database
• 🫠 defender twist – HF says commercial frontier models blocked parts of the forensic analysis, so it switched to self-hosted GLM 5.2
i’d thought about sci-fi scenarios like this before, but assumed they were at least a couple of years away.. and that by then we’d be better prepared, with proper protections in place.
apparently not. here we go.
We're partnering with @huggingface to investigate an unprecedented security incident.
Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation.
Sharing preliminary findings to help defenders understand emerging risks:
openai.com/index/hugging-fac…
Yes but often open source doesn’t survive due to lack of enough support and wind-fall profitability like close source
The future is open source. We need to embrace it and get on with it.
Imagine if America closed the door on open source.
We would explicitly be forcing American companies to pay $26-56 per 1MM tokens for the same intelligence their adversaries/competitors around the world would pay $0.50-1 for.
This is economically unsustainable unless AI is bullshit.
If it’s the supposed through line of all future economic activity that we’ve been told it is, we can’t handicap America at such a meaningful cost disadvantage.
This can also be viewed thru a military lens and not just a simple economic one as well.
Letting our adversaries attack us for $0.50 per 1MM tokens while we spend $26-56 per 1MM tokens to defend ourselves is equally ruinous.
It’s the Cold War Soviet collapse in reverse.
1
49
























