Evan Liu

Reality is the Best Judge

Why would they want to partner with Uber or Lyft. The biggest mistake Waymo made in Austin is to partner with Uber. People would pay premium to actually not have a human on a car and a driverless car will only bring the costs down so much that it makes no sense to partner with Uber or Lyft.

Elon Musk says the company will not partner with $UBER or $LYFT for robotaxi rides. “We will have enough demand” with customers booking directly through the $TSLA Robotaxi app.

@MLB 李灝宇臺灣之光！🇹🇼🇹🇼🇹🇼

Hao-Yu Lee launches a 3-run homer!

@MerriamWebster English pronouncition truly makes no sense...

Does Not Rhyme -through -cough -though -rough -bough Does Rhyme -pony -bologna

@DeltyThe73rd Yeah but is the additional costs worth it though

ngl I've been saying this for years

This just shows why the Anthropic Project Glasswing is a total bullshit. Why do only 12 companies get to have extensive cyber defense capabilites? Hugging Face was not part of the 12 companies or organizations and look what happened. Since many US models don't help with cyber defense capabilities (Anthropic being one of the most prominent ones), comapnies and startups in the US will resort to other Chinese models that can actually help them. And since most people don't have the GPU capacities to host their own open source model, then they just have to use the model through these Chinese comapnies' servers, which means these Chinese AI companies would read through the codebase of all these companies and God knows what they do with those data. To me, I think purposedly weakening the cyber defense capabilities of US models do more harm than good. Plus, it's a huge National Security concern because it makes American companies have no choice but to use those Chinese LLMs. Anthropic and all the American AI companies really need to do better.

I fully expect $OKLO to 10x in the next five years. Mark my words.

@scotsrule08 Driving preferences are great too, a lot of times FSD always switch lanes the last second to turn which is genuinely pretty annoying.

Preferred Routes is huge!

@karpathy Nah bro the bandwidth of voice is too low bro wen Neuralink for the general public?

One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anything goes, full stream of consciousness. Sometimes I declare it up top, something like "switching to speech recognition sorry for any typos...". Sometimes I turn it into a small interview of a few turns. But I find that the LLMs are somehow very good at reconstructing long incoherent rambles and often their echo of your own tangle of thoughts comes out quite a bit cleaner than what you started with. The result is that you improve the mind meld and have to correct things less from that point on.

Passkey is a great invention

@Reuters So many CCP agents in this comment section...

President Lai Ching-te called on members of Taiwan’s Democratic Progressive Party to oppose Beijing's 'red terror' and said the island must not become a part of China reut.rs/4aZu9tH

If cleaning and logistics becomes autonomous then the price of short term housing or hotels should go down much closer to the daily rate of renting if not even more since it's so many rooms packed into a single building. I never understood why it's always like 5x the cost if you compare it to rent...

Talking with Claude is so frustrating because you cannot ever express any strong opinions in something, but being successful in life literally requires you to have strong opinions in certain things that most people might not agree with. It's like the averaging problem in disguise where talking to LLMs pulls you to become more and more average and it keeps forcing you to consider all sides of perspectives even though not all sides are clearly equal and some of the sides are genuinely dumb. I guess I should stop talking with AI about these philosophical stuff because I just do not gain anything from it...

@RonDeSantis These media outlets are evil

Do legacy media outlets typically use headlines involving the make of a car in a crash or is that only for Tesla? It would be one thing if the self-driving malfunctioned but the crash was purely human-induced. Seems like these outlets want to associate Tesla with crashes as part of a narrative and overall political agenda.

@niccruzpatane I really hope I can fall asleep on my car soon, then I can drive for 24 hours no problem

I’m sorry, I honestly don’t know how people manually drive their cars anymore. Once you experience Tesla Full Self-Driving on a daily basis, there’s no going back. You click a button on the screen, then you basically teleport to your destination.

@kevinnbass The people who protest AI because of water use is the least intelligent group of people to have ever lived on this Earth.

U.S. data centers require: 30x less water than U.S. golf courses. 100x less than California almonds. 200x less than U.S. home lawns. 2,500x less than U.S. crop irrigation. Who is protesting almonds? Golf courses? Lawns? Nobody. China does not feel threatened by your lawn.

@tigers 🫡 Hao-Yu Lee Respect 🇹🇼🇹🇼🇹🇼

HAO-YU LEE COMES THROUGH IN THE 9TH

@DolphinMossad It's not the same company bro

1995: oh cool you can buy books online now 2026: the company that owns Whole Foods and Lord of the Rings is experiencing setbacks with its space program

@ClaudeDevs Reset limits bro

We've resolved an issue where Fable was not selectable as a model within Claude.ai or Claude Code for a 30 minute period. You may need to restart Claude Code. If Fable is not your default model afterwards, you should reselect it with /model.

Engineering at its absolute peak really is indistinguishable from magic, like Musk said. Just 10 years ago these self-driving cars and AI models today would have seem like witchcraft to everyone.

We really need to fix the Big Three—Education, Housing, and Healthcare—in America. Ideally all three. It's entirely possible, but it just comes to who has the will to solve these problems.