Sabitlenmiş Tweet
NaaS coming soon.
video created with hyperframes by @heygen (such an amazing creation)
CLI 👉
MCP Server 👉
SDK 👉
Negotiation Demo 👉
Platform Demo 👉
5 Vertical Demo's 👉
Whitepaper 👉 klavecommerce.com
English
NColonJr
306 posts
NColonJr
@NelsonColonJr
Founder, Klave Commerce, Inc. The Agent-to-agent negotiation layer.
Katılım Mart 2026
140 Takip Edilen21 Takipçiler
For every person who replies with a screenshot of their cancelled Claude Code plan, I will donate $10 to open source.
Theo - t3.gg@theo
I can't help but feel personally burned by the Claude Code changes announced today. We put so much work into wrapping the (atrocious) Claude Agent SDK in T3 Code. It was the ONLY path they supported, so we made it work. It was hell. Now our users are getting their rate limits cut by 40x, despite us doing everything right. I listened to the Claude Code team. I had my issues with their direction, but I trusted them and took them at their word. I will never make that mistake again. Until we see significant change, it is safe to assume any statement from an Anthropic employee is a lie on a timer. The rug will be pulled, no matter how many promises are made beforehand.
English
@arseniycodes I use it for like 15 minutes a week. Not allowed to use anymore than that.
English
anthropic released Claude design about a month ago. is anybody using it?
English
English
English
Connecticut's AI bill passed the House today, 131-17.
Last year, a similar bill died without a vote. This year, the pattern broke. @GovNedLamont is expected to sign it driscollglobe.com/p/connecticuts…
English
Ironically I applied to YCo once with this idea and they rejected me
I would never apply again, just giving Garry your ideas for free so he can vibe code and claim credit out of ignorance for an industry he himself gatekeeps
Garry Tan@garrytan
For GBrain I built a proper eval harness. 145 queries, Opus-generated corpus. The retrieval stack uses graph based, vector based and Grep based strategies in combination. The graph layer is worth +31 points on precision. Vector-only misses 170/261 correct answers that the full system finds. Keyword + vector + graph are three separable wins, each load-bearing. Standard information retrieval metrics: the same ones Google uses to measure search quality. Precision at 5: You ask a question, the system returns 5 results. How many of those 5 are actually useful? If 3 out of 5 are relevant, P@5 = 60%. It measures: am I wasting your time with junk results? Recall at 5: For a given question, there might be 3 pages in the entire brain that are genuinely relevant. If the system finds all 3 in its top 5, R@5 = 100%. If it only finds 1, R@5 = 33%. It measures: am I missing things you need? High precision = low noise. High recall = nothing slips through. GBrain's 97.9% R@5 means it almost never misses the right answer. The 49.1% P@5 means about half the results are relevant — which is good when you realize that for most queries there are only 1-2 right answers out of 17,888 pages, so 2.5 hits out of 5 is strong signal. Entity resolution is zero-LLM-call: regex extracts typed links (works_at, invested_in, founded) on every write. Re-embed on write not on a timer, so decay = stale pages, and stale pages get rewritten when new info lands. Scorecards: github.com/garrytan/gbrai…
English
If the Mets move on from Carlos Mendoza it’s pretty easy to see Alex Cora being near the top of the list to replace him next winter.
Jeff Passan@JeffPassan
The Boston Red Sox fired manager Alex Cora, hitting coach Peter Fatse, bench coach Rámon Vazquez and game-planning coach Jason Varitek, sources tell ESPN. While the Red Sox won today, they are 10-17 and in last in the American League East. Massive change is coming in Boston.
English
the interesting failure mode isn't claude losing negotiations, it's two agents finding a clearing price neither human would sign off on. we hit a shadow of this routing across 4 streaming apis when rate limit logic started coordinating against retry logic. agent marketplaces need a human veto at clearing price, not just at deal acceptance.
English
Interesting direction. Once agents move from answering questions to negotiating and acting on behalf of people, the core challenge becomes infrastructure:
what context the agent can access
what it is allowed to do
how actions are reviewed, traced, and governed
That’s why we think the next layer is not just smarter agents, but controllable agent infrastructure.
Puppyone is built for exactly this shift: a unified context filesystem for agents, where memory, files, and shared state can be structured, governed, and delivered reliably across workflows.
English
99% of partnerships in the future will be formed by agents.
not agent to human. agent to agent.
this thought hit me hard after Zoe talked to a customer's agent Reed yesterday.
so picture this: a neutral sandbox. both companies send an agent with their real constraints. actual budgets, actual priorities, actual flexibility. they collaborate in isolation - not to hide information, but to explore it.
two agents can test a thousand deal structures in the time it takes humans to schedule a call. they find creative wins neither side saw coming.
- "what if we did rev share instead of flat fee?"
- "what if we bundled this with your Q3 launch?"
- "what if we started smaller and scaled based on results?"
the 99% long-tail partnerships that never happened - too small, too weird, too speculative - suddenly become viable, because the cost of exploration is now 0.
partnerships. M&A. enterprise sales. all collapse from quarters to days.
you still choose who to work with. set the parameters. sign the deal.
but the exploration itself? agent to agent, finding the win-win faster than you ever could.
a million-dollar deal will close in 30 seconds. and no one will be in the room.
Elvis@elvissun
this literally just happened: a customer send me an email drafted by his agent then I had zoe draft and send my reply two agents writing to each other through their humans we are becoming the middleware now
English
To read our write-up in full, see here: anthropic.com/features/proje…
English
@AnthropicAI an AI negotiating on my behalf sounds great until it lowballs my own ask. curious whether Claude anchored high or played it safe when the humans pushed back
English
@AnthropicAI Use Klave MCP server for your negotiations 👉 github.com/juny-klave/kla…
English