正在加载视频...
视频加载失败
Jev Engineering is the layer most agent stacks are still missing. state → decision → action → verification → next state up to 193x faster and 444x cheaper in our tests. everyone is optimizing the model. Jev Engineering optimizes what happens between model calls. every agent eventually hits the... show more
27 条评论
I’m sorry for people who really miss it
lol, me too
It’s really new meta Ricker, banger article
fact bro
Saved g!
a model that refuses to write prose, finally
yep man
When Jev loops?
what a visual Ricker
@0xRicker hitting the nail on the head with this. I found optimizing between model calls made a huge difference in my own builds too.
if decisions become their own layer, who writes the ground truth you verify them against?
optimizing the process between model calls is key to efficiency gains
the important split is state versus a decision. state tells you what happened. a decision needs a reason, source, and expiry, or the next loop just repeats yesterday's mistake.the important split is state versus a decision. state tells you what happened. a decision needs a reason, source, and expiry, or the next loop just repeats yesterday's mistake.
state transitions are expensive. routing every decision through the model is why latency hurts. what is the actual mechanism that cuts 193x?
the glue between model calls is where most of my agent bugs show up. the model itself was rarely the problem.
« up to 193x faster and 444x cheaper in our tests » from which model?
The control-layer thesis is right: most agent stacks waste a model call deciding what to do next. The real benchmark is whether verification catches bad state transitions cheaply enough to make that loop dependable, not just faster.
optimization layer is literally what agents need right now
Optimizing the overhead between model calls is where the real multi-agent scaling happens
Jev is an if-statement and a waste of time, stop deepthroating the hype cycle
Chuẩn luôn fen ơi, đợt này mình cũng thấy thế :))
把 state → decision → action → verification 显式化，确实比把所有控制逻辑塞进提示词更容易优化。尤其要把状态、预算和终止条件作为运行时数据记录下来；提示词只描述目标，系统才能比较不同策略的成本与可靠性。
everyone optimized the brain and forgot the nervous system
Jev engineering so popular now
Everyone's sleeping on Jev.
optimizing the gap between calls is the real claim not another model swap
Thank you very muchaa 👏
![this harness stack is f*cking gold TypeSafe AI + LangChain finally dropped the layer that makes agents affordable at company scale: jev running inside the langchain agent loop as middleware the problem every big team hits: each fork in the loop is another full LLM call. at 10 agents that's a bill. at 10,000 it's the budget > routing before: every task goes to the frontier model now: jev picks the least costly model that can finish it. lookups go fast, architecture goes powerful > guardrails before: the "is this command dangerous?" check sat in the closed parts of claude code, codex, cursor now: AutoModeMiddleware(tools=["bash"]). one line, any agent, blocked before the tool runs > decisions before: one LLM call per question now: every question about the same state in one request, answered in parallel > audit before: "the model decided" now: the answer, the probability and the confidence stay in agent state > adoption no rewrite. pip install langchain-typesafe and drop it into the agent you already run TypeSafe reports up to 200x faster inference and 400x lower cost than comparable LLMs on classification. the LLM still reasons and writes. jev takes the thousands of small calls in between already live: browserbase agents for fractions of a cent, a live trading agent, email triage at scale full loop mapped below: 7 agents, 4 jev gates per step count the forks in your agent loop tonigh](https://image.24vids.com/tw-2104876056893104461/amplify_video_thumb/2104866254146203648/img/J8LnvuFvwbGYaJ7N.jpg)
