Founder. Prev built firmware and software systems in fintech, miltech, and medtech translational research
Singapore
Joined September 2009
- Tweets4.5K
- Following928
- Followers435
- Likes15K
🎯 and this will happen for all industries as capabilities mature and adoption broadens
I'm starting to sense a divide in the tech industry:
- People thriving on autonomy who are reinventing how they work and what tools they use with AI: LOTS of work, but they're enjoying it and getting very efficient
- People used to being mentored / trained / or simply told what to do and how. Frustrated with the lack of guidance from management on what to do with AI; also frustrated by seeing some of their peers thriving; increasingly bitter about AI and the state of the industry, as a whole
Neither group seems to be aware that there's another group, or have empathy for the other group
Vanity metrics aside, 💯 code quality matters, write linter rules / custom linters for bad patterns, they spread like a weed in the garden
Happened before agents too tho at a much slower pace
Can also recommend a file per function, named accordingly, helps reduce context usage
here's how i shipped 2,500 PRs last month to production
this was originally supposed to be for Cursor Compile in London. i couldn't make it since i was livestreaming for Grok @Bot Galaxy so i'm making it available for free here on X! watch it on 2x speed, i talk slowly
Paul Martin retweeted
The eMudhra compromise is an absolute joke. This isn't a standard data leak it's pure, unadulterated structural incompetence.
8 globally trusted intermediate CA keys were generated with basically zero cryptographic entropy. 😠 A standard NIST P-384 private key needs 384 bits of entropy so it's mathematically impossible to crack.
Yet, the blokes at @eMudhra generated production keys with 3 to 8 bits. The actual private key values were literal single digits like 4, 56, and 115. This is playground-level stuff. Infuriatingly amateur.
How does a private company with zero fundamental grasp of public key infrastructure even get global root trust? Indian regulatory and auditing bodies make us look like complete idiots. They treat cybersecurity like a casual compliance checkbox instead of critical national infrastructure.
This failure leaves the entire internet with a massive cleanup job. Browsers blocked these keys fast, but millions of unmaintained Linux distros, embedded IoT tech, Docker containers, and backend apps are completely exposed out there.
If you're an everyday user (not a developer), you still need to fix this on your personal devices right now:
Update everything: Force-update your web browsers (Chrome, Edge, Firefox, Safari) and your operating systems across your laptop, Mac, and mobile phone. Don't wait for auto-updates manually trigger them in your settings.
Test your device: Go to emudhra.we-have-a-problem.co…
Verify the result: If your browser blocks the site and says "Did not connect" (or throws a certificate error), you are safe. If the website opens normally, you are still vulnerable.
If you're running Linux, Python, Go, or Node infrastructure:
Stop what you're doing. Manually update your ca-certificates bundles, pins, or drop a system-level blocklist immediately to prevent traffic interception.
The casual negligence of these agencies shouldn't be your vulnerability.
The Internet goes on high alert after discovery of 8 compromised TLS certs arstechnica.com/security/202…
🤯
Some stats on #php2shell. In the 10 hours before it was publicly announced, across the @Cloudflare network we saw thousands of attempts to exploit it. That jumped to tens of thousands in the 45 minutes post announcement, hundreds of thousands within 2 hours, and millions now.
Holy shit the engineering and control required for that landing! Amazing safety system, but how many gonna be driving off cliffs & bridges knowing they'll land safely to get views and likes 😭
1980 vibes with Conway's Game of Life, some observers thought of the pixel clusters as if they had will, intent, and agency
I don’t understand why this consideration of “being kind” to a model even enters the calculus.
Are you “kind” to a calculator? Are you “kind” to your remote control?
I’m sorry but allowing emotion to enter your mindset when interacting with a model is a symptom of AI psychosis.
And to be clear, I’m not saying “cruelty” is good either. In fact, it is just as bad because, in both instances, you are attributing human emotion to an inanimate object.
A model is a thing. You simply use it. There is no “kind” or “cruel;” there is “productive” and “unproductive”—that’s it. Anything more is allowing your mind to get twisted into believing a weird fairy tale.
This is how it begins, if you thought echo chambers on social was bad wait till you see echo chambers for AI/SI
Empathy for humans should be #1 if human is cursing it most likely means the model is wrong, easy cue for bug fixing be it fine tuning or deeper training data
The quality bar of a $5T company 😱 what kinda progress bar is that, and those white lines at the top and bottom of the window buttons
😮😎
Starship transits the sun on its 14th flight.
It is now orbital.
Me for @WeAreSpaceScout
Paul Martin retweeted
This is so cool. DHH absolutely roasting the frontier-lab doomer class, at Rails World 2026 in Austin.
taking a sledgehammer to the “I worked at a frontier lab, therefore trust my apocalypse forecast” mindset.
"Maybe we should lean in with a little bit more optimism, with a little bit of P-Bloom instead of oh so much P-Doom. The chances of this panning out and us getting abundance and joy is so vastly greater than us getting the doom."
----
From "Ruby on Rails" YT channel, (link in comment)
It's time to put down the chisels and embrace the JCBs of the trade, skill is still required
DHH is probably one of the top software artisans of the past decade. Hearing him speak with this much conviction and transparency about one of the most polarizing topics in software right now, writing code by hand vs. with AI, is something pretty much every software engineer should hear a few times.
Especially those still on the wrong side of history.
Room for improvement for sure but we must not kid ourselves, there is zero way to tell good AI from real life when viewed through pixels
If you think that's scary wait till neural implants are widespread we'll have the same problem even looking through our own eyes
💯
The "whiteboard defense:" I should be able to pull you aside at any moment and ask you to explain any customer-facing system you've shipped. You should be able to clearly explain how it works and defend the decisions you made. This is my benchmark for responsible AI usage.
I don't expect line-level familiarity with the code. I don't care if you remember the exact function name or implementation detail. You may not even know it. I don't care.
But if I ask "why did you do X instead of Y?", "what happens if this actor behaves maliciously?", "what data structure did you use here and why?", or "where does this fail?" you should be able to answer confidently.
For PoCs, demos, experiments, whatever: I don't care. Generate 100% of it and understand none of it. Speed over quality every time in those specific scenarios.
But if you're shipping customer-facing work, you can't be shipping things you don't understand at a high level.