Hunch maps an accessibility snapshot to the next browser operation with Jev, then verifies outcomes in code and hands uncertain or irreversible steps to an LLM.
Hunch maps an accessibility snapshot to the next browser operation with Jev, then verifies outcomes in code and hands uncertain or irreversible steps to an LLM. The author reports 153 ms median decisions, 24/24 correct, and a four-step form fill in 3.4 seconds.
Breaking: Browser Use + Jev = Ultrafast ⚡ Findings flights took 7s and cost only $0.0039 🤯 > new action space every step > DOM state space > small LLM fallback to type (this video is at 1x speed btw) Built a tiny open source browser agent. try it below ↓
@gregpr07 · Agents & browsers · ~$0.004 · ~7 s
@gregpr07
X post
Agents & browsers
~$0.004
~7 s
just shipped something that made me go "wait, what" added Jev to my coding agent loop. one job: answer "is the task done?" that's it.
@cha73066 · Agents & browsers
@cha73066
X post
Agents & browsers
-
-
I just created Jevmind 🧠 Watching AI agents make decisions drove me crazy. They write a whole paragraph, then guess what it meant. No idea how sure they are no memory of being wrong. So I built a brain for them: → every choice becomes a clear answer with a confidence score
@dealerdefi · Agents & browsers
@dealerdefi
X post
Agents & browsers
-
-
Releasing Jev Desktop as open source. This is about 10x faster that GPT-6-Astra's native computer use in limited testing. This will be maintained as it's planned to become a core part of my internal verification workflow. https://t.co/9rFzo3HFnc
@Divine_machine · Agents & browsers
@Divine_machine
X post
Agents & browsers
-
-