I barely used twitter last 1.5 years. Today I open it, and I have hundreds of spam messages in my DMs. Way to go twitter! DM Open was a wonderful way to meet interesting people. An API for experts with deep knowledge that you could actually contact.Not anymore? DM open means spam

Pero en 2006 estaba y se sentía muy mayor. Tenía 84 años y la apariencia, exagerada por su prosa, de una iguana. Así que declinó la invitación sin perder la oportunidad de marcar, aunque fuera un poco, las mentes y los corazones de los jóvenes estudiantes neoyorquinos.

One advantage of using chatGPT instead of search engines is that it helps me preserve attention. It brings you the answer to your question, nothing more. SERP are chockfull of potential interesting departures from my straight path of thinking

Evolutionary tree of LLMs. Source: Yann LeCun on linkedin

Twitter doesn't help thinking clearly in that it violates these two recommendations: - **Don’t follow the crowds, but instead think for yourself.** * **Don’t make decisions while angry. (On twitter, anger is a very common emotion)

Any platform that optimizes for engagement will have this problem, as anger is the one shortcut to engagement these companies have found -also gaming companies So in general, social media will have a strong emotional component (negative emotions) that don't help thinking clearly

This is a very impressive new open model You can try it out here: huggingface.co/spaces/uwnlp/g… It passed my "sentient cheesecake that teaches people SQL" test with flying colors!

Will Russia station at least one nuclear weapon in Belarus before 2024? My prediction: 99% likely. metaculus.com/questions/1562…

Check out my latest article: New trends like LLMs and Apple Vision may make linux/BSDs harder to justify for productivity linkedin.com/pulse/new-tren… via LinkedIn

Today I'm giving this talk at pydata Berlin around 7:30pm, join me if you are curious: Debugging your reasoning (and code) with LLMs meetup.com/pydata-berlin/…

This was my #1 learning from my time at OpenAI. If someone else can't build on top of your work due to complexity - it's likely just a hack and not a fundamental step forward

DSR batch 36 demo day Tomorrow Tuesday at 5:30 pm at #SAP. Drop by if you have the chance! Data Science Retreat presents 8 Machine Learning prototypes and projects by Batch 36 participants. The event is free to attend thanks to #sap Berlin for hosting us! meetup.com/data-science-r…

In 1930, the mathematician Frank Ramsey proved that as graphs get bigger, structure is inevitable. Ramsey died later that year at the age of 26. This work spawned Ramsey theory, which looks for inescapable patterns in a huge range of systems. quantamagazine.org/after-nearly-a…

The current state of the EU's Cyber Resilience Act (CRA) will have severe and negative consequences for open source communities. Help us #FixTheCRA by making your voice heard. Learn more at linuxfoundation.eu/cyber-resilien… Linux Foundation Europe @ep_industry

Este video debería tener un millón de RT: Cayetana demolió en solo 2 minutos el verso del lenguaje inclusivo y el populismo. Subí el volumen y que exploten las redes. 🔥👏🏻

Being grilled by a devil's advocate forces you to refine your marketing messages, target customer profile, and feature set. There's no room for vague generalities or dodging questions you don’t have good answers to. That’s why it’s so useful.

I don’t grow and change as fast as I did 20 years ago. Maybe that’s because I’ve accumulated so much knowledge and wisdom, there’s less space left to grow in. Or maybe because I’ve ossified into a static attitude and person. Probably depends on the domain.

Here's a list of ALL the bad ideas most indie hackers have - Platform to validate business ideas - AI post generator for social media - AI photo generator (too late guys) - NextJS boilerplate to build SaaS - Platform to exchange feedback - Directory of launch platforms -

"For my next magic trick, I will make an internet troll apologize." –Transparency (cc @levelsio)