Excellent, I can finish my @buzzdotxyz compatible "safety harness" for my kids 🙌🏼
we announced Jev 1.14 on 10.14 at 1:14pm PT 🔥
a new and improved model, along with experimental features for our dreamers:
• images
• a new box type
• longer context
surprise, Jev was already able to see you! 👁️
BUT our bar for production-grade intelligence is quite high, so builders beware, these are preview features.
enter the ~*~ danger zone ~*~ at your own risk. 🚨
Aaron Alaniz retweeted
introducing . . . [ SQUARE KIOSK ]
massive display, outrageous industrial design, integrated payments, and a DEEP intelligence layer
truly a beautiful piece of hardware!
of course, something only @Square is capable of
ONWARD -----> so proud of the team for this one
Aaron Alaniz retweeted
Buzz Desktop v0.5.24 🐝
Give each agent a shared channel conversation or separate conversations for each thread. Plus clearer setup, remembered Inbox filtering, better startup recovery, and fixes for agent avatars across communities.
github.com/block/buzz/releas…
A week into working with buzz as my primary agent harness and it's funny watching all the agents think I'm fluent in git SHAs
Berd is a great harness that I used internally for a couple months as a Codex alternative. I really appreciate the first class representation of agents and this paradigm is serving me well as I explore Buzz more, but I still use Berd for certain flows. Congrats team!
We often build the tools we need ourselves. Berd is the desktop application our teams use to work with AI agents across projects, skills, tools, and models. Today, we’re making it open source and sharing what we’ve learned. Meet Berd, AI with character: block.xyz/inside/designing-a…
Buzzing after the latest @AcquiredFM episode, I re-watched Lightyear with my kids and wow, the scene where Buzz is falling out of the ship bay into space is incredible.
I was trying out Cerebras a few weeks ago and discovered and patched a thinking bug with @goose_oss
Released in 1.46.0. So fast!
github.com/aaif-goose/goose/…
I switched to Buzz as my agent harness for a couple days. I see it. A few hiccups, but this is the way. The agent-to-agent coordination works wildly well.
And as a bonus, it is surprisingly fun.
Aaron Alaniz retweeted
I wrote about the completely wild incident where OpenAI were testing a new model and it broke out of its sandbox and broke INTO Hugging Face to steal the answers to the benchmark simonwillison.net/2026/Jul/2…