If all you have is a hammer, everything looks like a nail. AI and agents @atlanhq. Solving for climate https://xitter.cf/t.co/gFMr60HnY0. thoughts are my own

roaming
Joined October 2021
We're planting trees wrong, and it's costing us billions! Up to 80% of planted saplings die because they're in the wrong place. We're fighting climate change with one hand tied behind our back. So we fixed it. 🧵 (1/4)
2
3
9
3,669
Everyone is always rooting for you. Your parents want you to be a great son. Wife wants you to be a great husband. Your boss wants you to be a slam dunk hire. Every first date you’ve ever been on they’ve been rooting for you to get laid. Every time you started to tell a joke people hoped it would have a hilarious punch line. Your proximity to anyone is a reflection of themself, meaning the deck is never stacked against you, and your failures are completely your own
Readers added context they thought people might want to know
This post copies word-for-word a May 2026 post by @AdemLuz. x.com/AdemLuz/status…
246
1,871
103
31,712
2,773,684
does anyone still prefer hermes over instinct? curious what use cases hermes genuinely outperforms instinct in for you?
53
For the last 100 days, I've been living Groundhog Day. Same questions about context. From CIOs. CDOs. Engineers. My mom. So I did some matchmaking. The people who've built it will answer live at Context Conference, Oct 28.
8
18
6
90
13,039
WhereToPlant is now part of claude startups 🌱 We're going to cover whole of India in next 6 months!
5
1
10
413
Not looking to earn profits from this initiative.... but amazing to see all the recognition this project is getting! you can just do things!
3
95
Rishav retweeted
Fifa Street PC Port for Macos available now!! github.com/fyzanshaik/fifa-s…
Apple silicon MacOS (arm64) build available now: github.com/fyzanshaik/fifa-s… Reused the existing Vulkan rendered for windows support by translating all Vulkan calls to Apple's metal using MoltenVK. Operates at 24.6fps performance improvements not a goal now, this is playable!!!
2
3
1
10
1,236
pplx-decider-v1.1-27b scores the highest on Decision Bench for accuracy, with the lowest cost. It reached 94.5% accuracy across 1071 cases at $0.017 per 1,000 decisions. Full evaluation: decisionbench.ai/ Get Started: pplx.ai/decisions
21
25
11
267
21,541
maybe we humans were writing the real slop
This quoted post is unavailable.
1
2
128
they nerfed opus
2
189
Perplexity Decider ranks first on DecisionBench (and lowest cost too)
Replying to @perplexity_ai
@perplexity_ai’s Decider V1.1 ranked #1 on decisionbench.ai with lowest cost! 949 shared text cases: 93.9% accuracy · 534 ms median · $0.016 / 1k decisions Thanks @AravSrinivas & perplexity team for building - looks really good, we're going to try it for our product use cases Our latest results live on decisionbench.ai xitter.cf/rohanatlan/status/2103…
46
32
5
327
41,035
Replying to @perplexity_ai
@perplexity_ai’s Decider V1.1 ranked #1 on decisionbench.ai with lowest cost! 949 shared text cases: 93.9% accuracy · 534 ms median · $0.016 / 1k decisions Thanks @AravSrinivas & perplexity team for building - looks really good, we're going to try it for our product use cases Our latest results live on decisionbench.ai xitter.cf/rohanatlan/status/2103…
Made decisionbench.ai to test because I wanted to see how well Jev actually performs compared to other models. Tested 12 models across 35 tasks. Jev did really well for how little it costs. Wrote up what I found 👇 View article
1
5
2
31
36,030
MCP stack is maturing! finally 🙏
Today, we’re launching M3, The MCP testing framework. Free and open source.
1
4
141
crazy good!
Replying to @sineframeHQ
We know that testing MCPs with real agents and harnesses is a pain, and M3 helps you make that process seamless. Just give your agent the M3 skill, and see it write tests using M3, tests that work across Claude code and Codex, any version.
1
1
57
Wrote about growth, trees, and Instinct and Muse growing like weeds, but in very different ways
1
3
9
781
how do i get a cosign invite?
1
116
12
131
7
4,403
67,779
i can't believe how bad LLMs are at estimating execution time. No it won't take you 2 days, You're gonna finish it before i finish my coffee
6
144
here we go!!
Made decisionbench.ai to test because I wanted to see how well Jev actually performs compared to other models. Tested 12 models across 35 tasks. Jev did really well for how little it costs. Wrote up what I found 👇 View article
2
131