Your customer’s agent is coming to your business. Coding agents are outrunning review. Researchers found agents can know they shouldn’t act, then do it anyway.
Plus AI in Korea’s bank hacks and Atlassian giving agents identities.
Daily Agentic is up. https://t.co/vxjJWwWJpk
So, they had us at "Le Chonk" -- we admit it. Clearly, a naming agency was not engaged for this assignment and we respect that. Lots more in this edition, from a new personal agent protocol to a new tiny Gemma from Google to a Black-Box viral sensation. https://t.co/fd5dCMRj3U
TikTok gets the agent drop, from "want" to "got" faster. Meanwhile, sightings of fleets of agents executing complex intelligence tasks autonomously hark the down of a new era. Plus German AI and a new AI Box so all your context belongs to you. https://t.co/vvscnsQhZv
Here’s everything I learned about building Glass Town, my software factory:
• Why I put my software factory in a persistent env
• How to use an MCP server to manage the env
• How I used Entire to make the factory transparent https://t.co/twhhy0veIL
A prompt bloat flag at C.H. Robinson traced back to a code error that kept a service in a retry loop, using tokens again on every pass. Jason Ward on tracking AI spend with observability metrics.
The web was built for humans.
Agents said cool, but where’s our interface?
WebMCP, MCP Apps, and A2UI are building the second layer of the web — made for agents, with humans still in the loop.
@ASalkever breaks it down: https://t.co/mvvn6YL3Bg
Want to go deeper? Catch the conversations at AGNTCon + MCPCon North America, Oct. 22–23 in San Jose: https://t.co/lsAfUAg4Qi
The latest Daily Agentic -- a robot that can't give you the (pinkie) finger, two open decision models from MSFT and CF, an AI-generate film heads to theaters in China and more. https://t.co/wArT56zLs7
"As a company we are spending $2 per million token. What does it mean? It doesn't mean anything." Kuntal Patel and Abhinav Lad of Palo Alto Networks on why per-token cost is a weak KPI, and what they measure once agents start looping on their own.
Good lord...@Dpbrinkm can edit! Now i'm even more jazzed to be speaking at this.
Come join him, and me, @dsp_, @techgirl1908, and lots more @AgenticAIFdn's AGNTCon in San Jose.
Oh, and DM me if you want 10% off, I have a code!
We are excited to announce we are partnering deeply with Muse to enable agentic checkout with Shop Pay on all Shopify stores, offering people an easy and delightful way to shop and check out with Muse.
@hey_amandam I have been using Instinct and its amazingly good. Love that its primarily SMS. Does not talk to you (yet) but in terms of spatial awareness and general quality of interactions and proactive ability, its been quite useful.
@maddiedreese@OpenAIDevs At our conf this last week, one presenter talked about reverse engineering an entire vending machine POS via the API with Claude IN A DAY because they needed to flash it for a sim eval setup. Crazy world.
@trq212 Hey @trq212 want to write a blog on this for the @AgenticAIFdn about the decision and what it changes? We'd love that. Agents.md is a foundation project.
Xiaomi livestreams frontier-model RL, OpenAI adds more agent failure reports, Qwen and GLM modify the systems beneath them, and Google Home opens up to third-party agents. https://t.co/JTP6uOZtIr