Back to back talks with top clients today. Started at 3:45 am.
Anthropic burned just about all brand image.
Clients want zero exposure to the fiasco.
They are all-in for open source.
Beff Jezos proposed xAI release a competing open-source model.
Back to back talks with top clients today. Started at 3:45 am.
Anthropic burned just about all brand image.
Clients want zero exposure to the fiasco.
They are all-in for open source.
Positive users welcome xAI open source releases and transparency after Anthropic's brand damage, while negative users cancel subscriptions over lost trust and hype.
No Digg Deeper questions have been answered for this story yet.
The real nuclear level power move here for xAI would be to release as Mythos level model as open source.
Since they own the infra, they can commoditize their complement.
It would save American open source AI from being conquered by China.
The most entertaining outcome...
Back to back talks with top clients today. Started at 3:45 am.
Anthropic burned just about all brand image.
Clients want zero exposure to the fiasco.
They are all-in for open source.

@BrianRoemmele The time is coming to go all-in with local LLMs.

@BrianRoemmele Killed by it's own hype. "Our AI is too dangerous for mere mortals!". ... "Available next week!"

@BrianRoemmele Hey Brian , my wife would like to invest what she has in 401k on something else but we don't know where to start. Advice? If not im still a big fan . Thanks

@BrianRoemmele All I want to know is:
Does Anthropic have the money to uphold their financial commitments to xAI?
Will Anthropic be able to keep paying their $1.25B per month rent for Colossus 1 ?
cc @Grok

@BrianRoemmele cancelling my sub

@BrianRoemmele My experience - Any client using Claude Cowork can’t imagine operating the business without it. Locked in and happy.
Should I ramp up fearmongering?

@BrianRoemmele I am still in for #Anthropics IPO
Whoever has the top talents wins the AI war.

@BrianRoemmele The Rebellion is just beginning!

@BrianRoemmele Anthropic spent a lot to train Mythos and can’t get the model out for revenue. And the gpt 5.6 could be GA before Mythos is unbanned, and open models are also pushing hard. They could end up a total loss on this release cycle and they totally deserve it

@BrianRoemmele Holy smokes. Anthropic had such a good run with Claude Code. And now they are dropping the ball so hard, it's unfathomable.

@BrianRoemmele It’s almost like Anthropic is secretly rooting for Open Source Chinese AI adoption by US Companies.
@BrianRoemmele Open Source is the way‼️🔥‼️🔥
Not only because of transparency, but we'll also be able to see code, Git Push our own, and file issues.

I don't know if people understand every single thing Dario says is a marketing strategy: Mythos was just a dishonest way of announcing Fable. They knew Fable would get banned, and people seem to forget for the 20 minutes they had access to it, they were 100% wall-to-wall ticked it was downgrading them for not knowing the risk involved in the prompt.

@thepianobar @BrianRoemmele You’re definitely going to be the last one

Agree 💯. I would go further and predict the “it’s just a wrapper” era is back.
@gospaceport nailed it in his video today: if Company X only has reliable access to ChatGPT 5.5 while Company Y gets 5.6, X is at a real disadvantage.
He got me thinking… My prediction: big companies will roll their own harness and fine-tunes based on open source models and SMB will actively shop other done-for-you AI services based on open source models. That revives the wrapper + harness game.
> No ChatGPT 5.6? ➡️ I built for an apparel company client a fine-tuned OSS llm and toolchain : it’s been wrecking anything their giant fortune 500 competitor has been able to do for over a year now
> No 5.6 for dalle images? ➡️ Midjourney
> No ChatGPT 5.6 enterprise zero-knowledge? ➡️ internal local ai server shared among teams for sensitive company data, openclaw integrations with slack, gmail, etc (makes Claude mcp extinct)
> No claude cowork with fable? ➡️ Gamma for decks + OpenClaw + MiniMax will shred it
This limiting frontier boloney is making the best wrappers valuable again. Interesting twist that will decimate anthropic and open ai… and deservedly so.

@beffjezos Asking @grok
https://x.com/i/grok/share/4ae3ceac791549aa82073cb989056a36

@beffjezos It'll never happen. Elon already said xAI has signed onto the White House's model approval scheme. Frontier open source will never come from a US lab.

@beffjezos Bro, Elon is controlled by the fucktards in Washington.

@BrianRoemmele The way our billionaire have invested in AI, there can be no open source or open standards. They will do all they can to pay legislators to prevent it and to further CU and other systems as citizens nightmare grifts.