amavashev

5 ms·

1.
by amavashev 5mo ago
True website is not for you and in the age for AI is not even for people. Its for AI agents reading your website and deciding what to do with it: recommend it, skip it, integrate with it, etc.
2.
by amavashev 6mo ago
The no proxy approach makes sense for LLMs calls. The gap is non LLM calls. Often times damage is done by non LLM calls -- tool calls like sending email, add records, files, placing order, etc. Budget enforcement at the LLM layer wont work
3.
by amavashev 7mo ago
Agree, you need to your own code review, although as AI gets better, this problem will most likely be solved.
4.
by amavashev 7mo ago
The core argument here is that autonomous agents will need a economic envelope enforced before execution not after. This means agents must follow this flow: reserve, use, commit or release. Interesting how others are handling runaway agent
5.
by amavashev 7mo ago
technical breakdown: https://dev.to/amavashev/i-burned-153-in-30-minutes-with-an-...
6.
by amavashev 7mo ago
Demo (no API key needed): https://github.com/runcycles/cycles-runaway-demo
7.
Show HN: Cycles – hard limits on agent actions before execution(runcycles.io)
1 points by amavashev 7mo ago|2 comments
8.
by amavashev 7mo ago
I'm the author. A few design decisions worth explaining: The model downgrade is the most useful feature for daily use — when budget drops below a threshold, the plugin silently swaps claude-opus to claude-sonnet, gpt-4o to gpt-4o-min
9.
Show HN: OpenClaw plugin – hard budget limits for agent tool calls(github.com)
1 points by amavashev 7mo ago|1 comments
10.
Show HN: RunCycles – pre-execution budget enforcement for autonomous agents(github.com)
2 points by amavashev 7mo ago|0 comments
11.
by amavashev 8mo ago
Per-key isolation + model locking is a solid baseline — especially for multi-tool stacks where one shared key hides everything. One thing we’ve noticed though: spend caps stop damage, but they don’t prevent pathological behavior. By the tim
12.
by amavashev 8mo ago
Drift correlating more with constraint tension than raw step count matches what we’ve observed. Your external gate instinct is right, but the gate has to be structurally external, not just logically external. If the agent can reason about t
13.
Show HN: Scalerx.ai – Create, train and deploy AI agents(scalerx.ai)
1 points by amavashev 2y ago|0 comments
14.
Create, train and launch personalized AI Agents(scalerx.ai)
2 points by amavashev 2y ago|0 comments