jag retweeted
Muse Code can now generate 3D models!
Create textured models that you can print, use as game assets, etc. The detail is insane.
It's now live in the CLI
agentic commerce x tiya dinner by @adisingh highlights:
- grok bot was a favorite agent amongst the table
- mass adoption is coming through muse, a product that’s passing the mom test
- discussed the impacts of stablecoins and micropayments on agentic commerce
- agentic banking is becoming more relevant for businesses and individuals
- the best mutuals are found through X
- someone needs to organize a founder trip to sausalito
jag retweeted
“if Mark Zuckerberg knows he can’t get 2%, then it’s going to ads” $META muse @eric_seufert well said.
Sooner or later, hopefully sooner
jag retweeted
I think the partnership between @alexandr_wang and @finkd could go down as one of the most consequential in tech.
Alex brings a new generation of talent, fresh perspectives, and a culture of innovation to Meta. Combine that with Mark’s experience, leadership, and ambition, and the potential is extraordinary.
Very bullish on @Meta’s future.
jag retweeted
Meta reaches 3.64 billion people with its apps every single day. That's almost half of the global population
They are a huge beneficiary of the commoditization of AI models, using AI agents at scale and AI agents taking the efficient market hypothesis to the extreme
Bullish
I bought $META
My thesis is that they have one of the highest estimated installed compute, data center and power amounts versus their market cap so I think they're well positioned if new AI builds slow while demand holds up
They can use their installed compute to serve Muse at scale for retail
Meta is a direct beneficiary of open source AI at scale and the world is not slowing down (China is 75% innovation vs distillation imo - see Delphi's China Report) and Meta can benefit from improving open models and distribute AI through its apps used by 3B+ people
Zuckerberg is a chameleon and can fit in with any president or political party and I see that flexibility as a political advantage
Google is close on power but I think their search business is eroding sentiment and they're balancing frontier AI with protecting search economics and they've lost significant AI talent (Jeff Dean, Demis's time)
AWS wins on power in my estimates but its valuation is not as interesting to me here but they are a big beneficiary of open source model deployment. Meta’s strongest case is the combination of lower valuation, estimated installed AI compute and space per dollar, and consumer distribution via Muse. I'm not bullish Amazon Alexa.
I also think @alexandr_wang has the rizz to onboard mass young consumers to muse
I do like a barbell strategy where I'm long META, which has substantial installed compute relative to its market cap right now, and long SpaceX for the long term potential of compute in space.
Barbell of most installed to market cap now (Meta) and later (SpaceX)
I got really close on Palantir because it's the full encapsulation of my AI sovereignty thesis via a public stock: controlling your compute and data, using forward deployed engineers to transform a company, and leveraging open source models. I think that they'll probably do well, but my concern is on the valuation multiple right now and can be right and lose money as the multiple contracts
Not financial advice and only my thoughts. I will adapt my views as I get more information.
jag retweeted
Replying to @PhilipQiu5
Not just that, integrated UGC has way better conversion rates than VTO. Muse has such an advantage to build that into the agent shopping experience that no one else does!
jag retweeted
Muse launched Virtual Try On yesterday.
There's a pattern in agentic commerce of not learning from past mistakes. VTO has been tried many times before and has not had any user stickiness (Doji being the highest profile example). A few reasons:
> The bar for a good virtual try on is high. Fashion is a high consideration category and the UX needs to be creative and perfect.
> Visual based shopping is incredibly challenging. For AI form factors to take off, platforms need enourmous swaths of product data and user context.
Ultimately, the north star metric we should be focused on is conversion rates. Agents are still far from increasing conversion rates compared to existing shopping flows.
There is not enough discourse around how we can accomplish this.





























