Adi (@adstastic) header
Adi (@adstastic) profile picture

building a walkie-talkie to any agent @ https://teapawt.net/t.co/3cW6bqGkAE | cto @ familiar | prev @apple

San Francisco, CA
Joined April 2009
Photo by Adi (@adstastic)Photo by Adi (@adstastic)Photo by Adi (@adstastic)Photo by Adi (@adstastic)Photo by Adi (@adstastic)Photo by Adi (@adstastic)Photo by Adi (@adstastic)Photo by Adi (@adstastic)Photo by Adi (@adstastic)
Pinned Tweet
you pull out your phone to do one thing and end up doing something completely different. agents were supposed to fix that - now we juggle chats instead of apps. voice was supposed to fix that - now we walk around holding the phone up like a pizza slice. so we built guv, a walkie-talkie for your agents. hold the button, think out loud, it gets done. works with grok bot today. other personal agents and agents on your machine coming soon. have an apple watch? reply with the first thing you'd tell your agent for access.
We want our attention back. Launching guv.sh - a walkie-talkie to your agents. Hold the button, think out loud, it gets done. Reply with the first thing you'd say into it, we'll send you a testflight link.
6
6
2
26
767
Everyone has drawn this chart in their head and is making a generational bet on which line is true.
🤖 Made with AI
242
458
170
7,985
569,366
buildings can catch rockets now meanwhile i'm still cleaning up after my agents to avoid my drives filling up every couple of days
Mechazilla catches Starship! Congratulations to the entire SpaceX team on this incredible achievement!
25
thank you algo for the 🇯🇵 photographers in my tl feels like old internet to have these delightful niches
今まで長野で撮影した写真ぜひご覧ください。
4
224
enjoying delta.dev by @zeddotdev a lot i previously built terminal agent plugins for annotating agent responses, because they write SO much on every turn (even with caveman, etc). It's not claudish, there's usually a lot of context relating to a a simple question in a chat. delta makes the transcript a commentable surface, so rather than getting lost in text, i comment directly on specific words in the agent response, it responds to those comments, rinse and repeat until everything is settled. with the diff right there and able to mix in the discussion, it's my favourite experience right now for the ends of the work: locking in a design at the start, and reviewing a stack of PRs at the end
2
37
More CVEs…?
There has been a temporary telecommunications outage - out of an abundance of caution all flights have been halted. Every flight has been accounted for and international flights will be handled on a case by case basis.
1
84
the full rim or no?
1
38
glimpses of level 4 autonomy of personal stuff now. recent guv prompt: "late one tonight, change alarm if @janisveryklug ok to delay morning sync" what is this epic personal agent? claude code
1
2
98
qr codes are mundane but 3d printing them feels like alien script try it if you have a printer
1
1
64
great thread go read it but if tldr models are improving 2x faster than 6 months ago
Replying to @METR_Evals
1. Benchmarks provide one piece of evidence for acceleration. Since Oct 2025, AECI scores for Anthropic’s models have grown at 1.9x [95%: 1.5–2.8] the earlier rate. Analysis choices affect this estimate, and it remains unclear how benchmark scores map to real-world capabilities.
30
but it does write like old opus, which smells
some thoughts on fable 5.2 vs opus 5.5: the biggest difference is in how much attention/steering it takes to get things done. they talk similar, but fable is more pragmatic and actually steers subagents to finish stuff. biggest pleasant surprise has been it detecting when the loops aren't adding value and dismissing things subagents find. ultracode does a decent job of conserving fable limits - too early to tell but ultracode with sonnets might be the combo that makes the quotas last all week
22