We are going to come full circle and create a new language for prompt-code to better communicate with LLMs. The result will be a new reflective programming language that both humans and LLMs will learn and respond to.

So the employees at X Corp are called Xes?

This is exactly what the Doomers get wrong about AI alignment. The answer to AI safety is more training, not less. Poorly or incompletely trained models could be dangerous, so the answer is to train them on more data till you get it right. Go team Yann LeCun and

AI -> AR hype shift happening in 3..2..

Beware the Prophets and Doomers claiming the coming of false gods. Submit not your mind to such stories and lies of those who seek power over you.

Building smartphone apps was a ZIRP

So if we're on the verge of discovering superconductivity and AGI then it makes perfect sense that the aliens appear at this moment in human history, right?

Let the waters kiss And transmutate These leaden grudges into gold

Another one

Sam == Icarus

What Zuck is really saying is that the next Quest will have eye tracking and will be priced closer to $3000 now that Apple has normalized that price point

Legal team approves

Whenever someone says LLMs are AGI show them this interaction and ask them how many language tokens were predicted in this display of intelligence

Are Elon and xAI not involved in Stargate?? 🤔

Tremendous alpha right now in creating South Park memes

Fascinating example of fragility at the intersection of engineering complexity and human decisions. We live on the edge of chaos.

Everything is CRUD

The next step is to arrange agents in layers and train the agent network end-to-end with a unified and single loss objective

Agency precedes intelligence, not the other way round. You learn by doing.

Feel you're falling behind trying to keep up with everything happening in AI? You're not behind. You're early.