major tom retweeted
Today, we're opening up exclusive beta access to @nuehealth_ - India's first blood testing subscription.
We're giving 20 people free access to the Nue Health mobile app membership.
These 20 people will be the founding members of Nue Health and will have free perks for life.
How to apply:
major tom retweeted
Saw a guy call it cuck coding instead of vibe coding because you just sit in a chair and watch😭
Tell me you don't understand statistics without telling me you don't understand statistics.
This chart shows the share of today’s infrastructure “built after 2014.” But if an economy grows ~6% a year, about half of everything it has will always have been built in the last 12 years.
Take power. From 2002 to 2014, capacity went from ~105 GW to 249 GW, about 7.5% a year. From 2014 to 2026 it went from 249 to 540 GW, about 6.7% a year. The “54% built after 2014” headline hides slower growth than the decade before.
Coal went from 565 to 1,000 MT, about 4.9% a year, roughly the same pace as the decade before. Transmission went from 2.9 to 5.0 lakh ckm, about 4.7% a year. Gas pipelines show the same pattern. These are trend lines, not breakthroughs.
Solar isn't admissible here as module prices fell ~90% worldwide in the same period.
major tom retweeted
indian tech twitter is living such a delusional life that a single AI model announcement makes them tweet about it for days, yet they have no idea what’s happening in their own country outside their privileged tiny tech bubble
This randomly showed up on my YouTube feed and I’m really considering watching this entire course now.
Opus 5.5 is fixing my imposter syndrome.
The last couple generations of models did 2 things:
- caused orgs to prioritise shipping speed so much that engineers everywhere had large knowledge gaps about the code their models were shipping. Poor reviews, poor mental models, and essentially this state of constant anxiety due to being out of touch with things they’re accountable for.
- models had so much verbosity in their explanations that it wasn’t even possible to read whatever they were generating. It was impossible to ask an LLM about the code it was writing without confusing yourself even more.
Opus 5.5 fixed the second issue by being absolutely beautiful in terms of generating explanations using animations, interactive artifacts and just… very simple, easy-to-understand language. Not to mention, the code it writes narrows that gap even more.
On the other hand, orgs have slowly started to adopt better engineering practices after watching their software go through massive quality issues.
This feels like such a win. You don’t have to write code, reviewing is a lot easier, and you’re also now very familiar with the code you’re shipping.
major tom retweeted
we’re all just employedcels permanently deployed to jobistan until we are eliminated by osama bin laidoff


























