Mark Worrall

Started a substack: The Infinite Human [link in comments] Will explore the intersections of AI, knowledge, and education - topics close to my heart as I build an EdTech company. Also, want to pushback on some of this "AI is gonna takeover" narrative.

It’s been a year since I started Nodeledge. Some reflections 👇 - Most educational products are not serious about the science of learning. By serious, I mean willing to hold the line on what’s evidence-based and keeping a high-bar for learners, even if it means losing sales. If…

Why is typing on a phone keyboard still so bad? Presumably latency and storage / memory constraints are preventing innovation here. Still, whoever builds something that lets you type as fast as you can with errors and then transcribes to perfect text will be a billionaire.

The "theatre of learning" is a compelling illusion. We often 𝘧𝘦𝘦𝘭 like we're learning when we're merely re-reading notes or tabbing away in our AI-powered IDE. This phenomenon has been meticulously explored by the Bjorks (Bjork, E. L., & Bjork, R. A. (2011)) and highlights a…

Don't be fooled. This is a triumph of human, not artificial, intelligence.

Trying Cursor, currently going well. 😅

Whether you code with AI or not: you still need to solve people's problems. And speak to their emotions.

Remember: no-one is a prophet in their own town.

Most scroll. Few build. Active doing - not passive consumption - is the real edge in building your career capital. Your skills compound when you produce, not just consume. Yet most of us are idle consumers and caught up in the scroll epidemic. As Cal Newport notes, “deep work”…

Forget a citizenship test, surely the best way to know if someone's English is to see who's still having a hot cup of tea when it's 30°C outside. 🥵

Cheat on everything. Learn nothing. Don't trade your potential for dependency on someone else's product.

> At one point, Claude hallucinated that it was a real, physical person, and claimed that it was coming in to work in the shop. We’re still not sure why this happened. Because its training data probably contained scenarios or narratives where AI models are depicted as having…

As painful as it is to accept, when building a new product you need to be a pragmatist, not a purist.

Killer feature request: make Enter add a new line, not send the message. For the love of all that is good, at least let this be an option we can turn on. @theo

Sam Altman has just released a blog claiming we are past the event horizon and the takeoff to superintelligence has started. 😐🤦‍♂️ ⚠️ I’m a strong sceptic of the claims of a superintelligence / ASI. ⚠️ Nevertheless, I’m going to brave offering a few quick thoughts for the…

Might just call it a day and stop there.

My timeline is full of people bragging about vibe coding entire apps in hours. Meanwhile, I'm writing dozens of math and ML lessons almost entirely by hand (with some help from AI). Can't tell if I'm wasting my best years or playing a very smart long game.

As I keep saying: mistaking examples of reasoning for actual reasoning is a logical fallacy. Don't fall for it.

Back posting after a long-ish hiatus. Between newborn chaos and startup chaos, it's been hard to find time (or push past the cringe / tsunami of people blathering on about AI agents). But I'm aiming to be consistent now. If I fail, it'll be by going down screaming in a ball of…