James Guo

Talk with the world.

[James Guo @apacheguo]

For frontier model company it’s all about agi, or else it will end up into token price war.
For cloud company it’s all about ecosystem and cost per token, it’s service.

[James Guo @apacheguo]

We may get replaced by ai eventually in terms of labor. But that’s a process. What we can do is utilize ai to improve productivity first and use the outcome to buy ai assets.

[James Guo @apacheguo]

Two points from Peter Steinberger’s interview:
1.most app will disappear or become api
2.the code does not matter, it’s the idea and brand that matters.

[James Guo @apacheguo]

The company that provides intelligence at the largest scale with the lowest cost wins.

[Elon Musk @elonmusk]

@levelsio Hardware is the end game

[James Guo @apacheguo]

I happen to have the same habit with Sergey of talking to Gemini when driving. But he’s right, the current version sucks. Looking forward to the version he’s been using.

[James Guo @apacheguo]

Sergey’s latest lecture at Stanford corresponds to the point that how much google has accumulated all the way along the development of AI. It’s a process and culture, not a sudden change.

[James Guo @apacheguo]

Just watched Jeff Dean’s visit on x moonshot, it makes me feel very natural the whole process and breakthroughs he and google has been engaged on ai. For the next five years Jeff says he will be focused on making ai more available which I think goog really has a good chance to.

[James Guo @apacheguo]

For non-tech creators, ai removes the obstacle to build. For technical engineers, ai amplifies the productivity. Finally thoughts will be the only difference that define us.

[James Guo @apacheguo]

Spacex will start the new age of discovery, buy your tickets for the future.

[James Guo @apacheguo]

Communicating or collaborating with ai can be more efficient than humans.

[James Guo @apacheguo]

Joe Tsai’s four points on AI at HKU talk.
- top-down policy
- electricity
- talent
- open-source to accelerate ai adoption

[James Guo @apacheguo]

Just to mark this current point where fsd seems to be very close to get to the point to be a killer app. We should be able to see how this new tech creates new demands in 2026, and hopefully with an exponential growth.

[Elon Musk @elonmusk]

@DevinOlsenn 14.3 is where the last big piece of the puzzle finally lands

[James Guo @apacheguo]

The most unforgettable and happiness day! Daddy loves you baby!

[James Guo @apacheguo]

Kardshev scale is the real scaling law.

[James Guo @apacheguo]

@ramizwebti Maybe a new model architecture which solves the n square complexity of attention. In terms of accessibility google has deep vertical integration and also strong infra so I guess they will make the cost per token much cheaper.

[James Guo @apacheguo]

Just watched Jeff Dean’s visit on x moonshot, it makes me feel very natural the whole process and breakthroughs he and google has been engaged on ai. For the next five years Jeff says he will be focused on making ai more available which I think goog really has a good chance to.

[James Guo @apacheguo]

Obviously google benefits from the ai revolution which you can tell from the earnings report, no matter the GCP or google one or ai search mode. But the pe is constrained from the worrying of overturn by new companys like OpenAI. We’ll see in the next few years.

[James Guo @apacheguo]

The human brain is much more efficient at reasoning compared to the machine. But the machine theoretically can have unlimited computation source which makes it just a matter of time to exceed human intelligence.

[James Guo @apacheguo]

Maybe it will be grok who helped us to get to Mars.

[James Guo @apacheguo]

Current llms are a condensed expression of all the human data or knowledge which makes them very good at understanding language and helping you learn things. It’s a revolution of the way to acquire information and knowledge.

[James Guo @apacheguo]

But it’s still more a copilot in terms of research. What grok is trying to do to be able to seeking truth will really be the evolution of productivity, which will be the scaling of scientific research resource.