139 AI agents + generating $2mm ARR + 150.000 users Founder with AI agents running 98% of my startups 🎓 Harvard PhD dropout 🥂 + 1 exit
San Francisco, CA
Joined November 2025
- Tweets1.7K
- Following1.4K
- Followers389
- Likes1.6K
if you’re not marketing your product and doing great b2b partnerships for your company, you’re not a founder.
you’re just a dev
Definitely, GPT/Open AI is winning the race…
Anthropic hate for their clients has been too obvious
Today was the first time my AI locked me out of my own computer haha (this dumb Codex Sol Xhigh, by the way)
I could see my screen, but all typing and navigation were handled exclusively by the agents…
like only they could type and communicating with each other (I have a few dozen agents running autonomously + a loop here...)
Took me 2 hours to find a “back door agent” (one of my orchestrators) to ask to fix the problem…
Anthropic just made Fable 5 and Opus 5 worst
So now, they call it: cheaper + more tokens per week…
Both fable 5 and opus 5 got completely nerfed…
I really do not understand why anthropic scam people like this
they could just charge more for both… or even leave the Opus 4.8
I think they forced the switch to Fable 5 because it was cheaper and, obviously, worse.
Opus 4.8 was better.
Maybe even 4.6.
Actually I was wrong
Opus 5 is much more dumber than Opus 4.8
I am sure they will solve it in the next weeks, but now, not worth using Opus 5
Fable 5 or Codex 5.6 or Opus 4.8
My stack:
Opus 4.8 as orchestrator
Codex 5.6 Sun as auditor
Codex 5.5 builder
GLM and Kimi as workers
Fable 5 for strategy + front + copy
And of course my 2 best friends Hermes taking care of all my system
Just me or fable 5 started to be like Opus 4.8 today?
It is ultra dumb here (running a few billion tokens per week, it is not only 1 random chat/project)
Claude really like to play this marketing game…
Access for 2 days > ops, now we need to shut it down
After a few weeks > only 1 week access, enjoy as much as you can.
then > Only 1 more week
then, again > Only 1 more week
They think their clients are really stupid
best news of the year
Claude Fable 5 will be available again globally tomorrow.
After a series of productive conversations with the US government, we're redeploying the model with a new set of classifiers to target and block more cybersecurity tasks. In the near term, some routine tasks like coding and debugging will fall back to Opus 4.8. We’ll continue to refine these classifiers over the coming weeks to reduce false positives and better distinguish genuine misuse from legitimate requests.
We’ve also begun drafting a consensus framework—with Amazon, Microsoft, Google, and other Glasswing partners—for assessing the severity of AI jailbreaks and how AI developers should respond to them. We invite other industry partners and model providers to join us in this effort.
Finally, we’re scaling up our collaboration with the US government on model testing and safeguards. This will include pre-release access to models and safeguards for evaluation, information sharing on jailbreaks and misuse, and dedicated resources for joint research.
Thank you to our users for your patience, and to our partners across the government, industry, and the research community who worked alongside us to make Fable 5 available again.
Read our full blog: anthropic.com/news/redeployi…
This is one of the best practical ways to use loop
NEW Codex workflow:
1. Codex can have a thread manage other threads.
2. Codex can create heartbeats to keep doing an action or check in
3. Codex and use /goal on work threads to make sure agents complete all work
Using those two primitives, you can ask agents to schedule work, check in on it, address PR issues (Codex code review plugin), and push work forward until it's ready to merge.