Pinned Tweet
Great timing to announce that @CARTO just launched Agentic GIS:
Our next generation of AI agents, MCP Tools and features to solve and democratize geospatial analysis:
- securely
- at scale
- for boring enterprise problems.
Thread below ⬇️✨
speaking to my b2b heart: @gdb says “boring enterprise problems” are still massively underserved by AI startups rn
Tena retweeted
Visualising the vibe of ~11k restaurants, bars and cafes in Madrid!
The idea is to let Jev categorise places on an arbitrary scale that the user can define: “trendy” vs “traditional” or “ideal for a date” vs “ideal for a business meeting”..
Jev uses the name, location, description and a snapshot of the menu to judge each place. It’s able to categorise the whole 11k set in ~2 seconds (around 1m input tokens in total for about 4 cents per run). For comparison, even the latest Haiku 5.5 takes around 13s and $0.17 per run 🗺️
madrid-food-map.vercel.app/
This feels correct 👇🏻
Just this week I used two harnesses built by the same company with the same model (Gemini 3.8 Flash) and I thought one gave great results and the other was just… terrible.
Harnesses often get dismissed as just scaffolding, just prompt engineering, and not real research. But that couldn't be farther from the truth. The same model weights that score 30% on ARC-AGI score 95% with a better harness.
So we gathered a group of researchers and founders working at the frontier to do a deep dive into the state of harnesses.
We cover how we got to this point, the case for making your harness as expressive as possible, and what YC learned building an agent for every employee in the company.
00:00 - @FrancoisChauba1: Why harnesses matter
04:27 - Building an auto-researcher by accident
07:13 - A five minute history of harnesses
13:56 - Self-improving harnesses
18:35 - @sethkarten: Prime Agent, a self-improving RLM harness
21:50 - Context as an L1, L2, L3 cache
24:51 - From Turing machine to von Neumann computer
28:33 - Messaging between agents
30:04 - ARC-AGI results
33:09 - Emulator Bench and GPU kernels
37:30 - @JonSaadFalcon: OpenJarvis, personal AI on personal devices
38:26 - How far behind are local models
39:21 - The five primitives of a personal AI stack
42:47 - Letting cloud models optimize your local stack
43:53 - 800x cheaper than the cloud
45:58 - @josh__france and @jbellregan: QM, YC's agent harness for work
47:29 - A history of YC's internal agents
49:24 - OpenClaw and a fleet of 50 agents
51:04 - Pulling the brain out of the sandbox
54:43 - Letting the agent choose its own sandbox and model
57:16 - The grind tool: budgets on goals
58:50 - Agents don't understand social context
/goal Let me know when Github is fully back online. Check github status or whatever source. Notify me loudly somehow.
I feel targeted.
Cities: Skylines meets Balatro in Steam's coolest new roguelike in years polygon.com/combolands-demo-…
ALT Cities: Skylines meets Balatro in Steam's coolest new roguelike in years
Tena retweeted
A pre-commit hook for Git that reads in the diff'ed code on a PR and quizzes you on it before you're allowed to open the PR.
You have to score an 85% or above on the quiz to open the PR.
This would fix 99% of complaints about slop PRs overnight.
Tena retweeted
A slide from our last all hands.
I've also started to call out this behavior publicly in Slack channels.
Tired of consuming workslop
This diagram screams AI-generated with no edits.
Agents are part of a massive, interconnected ecosystem. But how do they find and trust each other across different platforms?
Today, we’re proud to announce the Agentic Resource Discovery (ARD), an open specification alongside industry partners (including Cisco, Databricks, GitHub, GoDaddy, Hugging Face, Microsoft, NVIDIA, Salesforce, ServiceNow, and Snowflake). ARD gives any agent a secure, decentralized way to discover and verify capabilities (like tools, skills, MCP servers, and other agents) anywhere on the web.
Read the full announcement and get started: goo.gle/4a2sTWf
ALT This architectural diagram titled "Agentic Resource Discovery" illustrates two parallel paths for an AI agent to find resources: a direct "Catalog" layer where companies self-host discovery files and a centralized "Registry" discovery service. The AI agent searches these sources, then verifies and connects to resources via A2A, MCP, or API protocols. A five-step process flow at the bottom summarizes the workflow: 1) Publish, 2) Crawl, 3) Search, 4) Verify, and 5) Connect.
Who is very responsive at @AnthropicAI?
We've been trying to get a response on our MCP directory submission and no answer for weeks.
I'd love to DM with details! 😔
Tena retweeted
Product management is the art of the writing the least amount of code for the greatest benefit to your users.
Tena retweeted
oh you’re using claude code? everyone’s using open code. just kidding we’re all on amp code. we’re using cline, we’re using roo code. we just forked our own version of roo. were using kilo code. we were on coderabbit but their ceo yelled at us so now we’re using qorbit. apple just acquired them for $30bn so we just migrated our entire team to slash commands. one guy is still on aider. the PM is on loveable. he just shipped a new product on replit. the intern installed a slackbot that lets you chat with your spreadsheet. legal is still reviewing devin’s enterprise contract. we evaluated junie for three ukrainians using jetbrains. someone in slack just asked “has anyone tried amp?” we are using goose for scripts. next week we’re piloting augment code. the CTO heard good things about trae. our CEO is friends with the guy from conductor. our CFO resigned. our CISO said we’ve had fourteen supply chain attacks in the last week. we’re shipping the worlds most expensive todo app.