Deepak Dhingra

honestly think elon pulled off the biggest comeback in ai and nobody's noticing codex, claude code, none of it is built for actual humans, I couldn't handle that weird infuriating communication style Grok for now manages claude code and codex and come back to me only where am actually needed instead of claude forcing me to watch every session as it needs a decision every 5 minutes of so it's a real chief of staff.

ChatGPT reminders is so underrated

I think the most underrated person in tech might be Priscilla chan

Instagram distorts your perception of your own life, Twitter distorts your perception of society

Ai is slop when someone else generates it but it's kino when you do it

Periodic reminder that people with literal mental disorders and a history of psychosis post opinions on social media. If you read a post that’s unhinged to the core, don’t make assumptions about society based on it.

Gen z should take notes, this is what elite level rage baiting looks like.

Matter of time before couple of firms will go bankrupt because of ai and that's going to send shivers down the spine of the whole market AI is going to be a covid level event and even though we have been talking about it for last four years, I think at the impact level we are currently at january 2020

Theories like Advaita Vedanta and others that claim to be anti-ego are, in reality, some of the biggest ego amplifiers. What can be more ego-driven than making people believe that they are the universe? The ultimate ego trip is convincing yourself you’ve transcended the ego because you’ve realized you are the universe.

The olympic committee should start preparing for a robotic category

Acceptance is a bliss

Most jacked guy in my gym is a 55 year old russian who watches TikTok on speaker between sets

People tolerate inequality more easily than unpredictability

What is hateful to you, do not do to your neighbour.

Someone finally said it, felt like a teaser for a tv show

So the PR for glm model stopped or what? It was the hottest thing before kimi

A strange possibility: as AI models like claude fable and others become more intelligent, they may increasingly communicate like scientists or philosophers, precise, nuanced, and technically correct, but difficult for the average person to follow. At some point, greater intelligence could make the product feel less coherent and less useful to most users. This may create a “dumbest AI wins” situation, where the most successful model is not the smartest one, but the one that explains things most simply and makes users feel understood. Optimizing purely for intelligence may become counterproductive.

The three in kimi 3 stands for three js it seems, all the showcased projects are webgl three js projects. We are out of public benchmarks now to showcase the quality of a product.