i will speak on behalf of my man. It was early days of our dating. I was on meds and was bleeding heavily. I got off from his car and I saw a huge stain on the seat (ps: i took all the precautions I can but still it leaked). I was embarrassed. I said sorry and said let me clean. He said - “no. I will take care of it. Don’t worry”. And he did took care of it. I said sorry many times. He said - “why? ! Neha, this is the part of life. It is just a stain.”

Introducing 𝗢𝗺𝗻𝗶𝗴𝗲𝗻𝘁, a meta-harness to combine, control, and share your agents. The best teams already mix models and harnesses and design loops that drive teams of agents. No single harness can keep up with that alone. So we built the layer above — we call it a 𝗺𝗲𝘁𝗮-𝗵𝗮𝗿𝗻𝗲𝘀𝘀. Omnigent sits above the tools you already use, Claude Code, Codex, Pi, and your own agents, and gives them one shared layer: • 𝗖𝗼𝗺𝗽𝗼𝘀𝗶𝘁𝗶𝗼𝗻: combine models, harnesses, and techniques without rewriting code, and switch between them with one-line changes • 𝗖𝗼𝗻𝘁𝗿𝗼𝗹: stateful, data-centric policies and cost budgets enforced at the meta-harness layer, not via prompts — let agents run without watching them • 𝗖𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻: share a live agent session via URL with full history, so teammates can review, comment, and steer in real time Every session is reachable from a terminal, the web, a desktop app, and your phone. We built Omnigent for our own use at Databricks and are open sourcing it under Apache 2.0. databricks.com/blog/introduc…

time to try for US citizenship to get access to Fable. I was sick and not able to try it once 🥲

A dev got so frustrated watching his AI agent write 500 lines for a 5-line problem that he built a fix. He called it Ponytail. Named after the guy every team has - long ponytail, oval glasses, been there longer than the version control. You show him fifty lines; he looks at them, says nothing, and replaces them with one. Now your agent does the same. Before writing anything, it looks for a reason not to. 80-94% less code. 47-77% cheaper. 3-6x faster. The best code is the code you never wrote. GitHub Repo: github.com/DietrichGebert/po…

100%...other wise the prompt would be just - build a mobile app with no errors. the person who has no knowledge of RN vs flutter vs native apps, DB, infra, orchestration, security etc...would always end up just sharing localhost URLs.

none..I am on Kiro :p (sorry for the bad PJ)

a few yes! I know a few but a lot are just loving the process and why not? as a human we need to be curious.

nah! depends..lot of engineers like me are using AI to speed up their ideas, deliver work and use experience to get best from the AI. non-engineers are using AI to get the taste of coding, building apps. in both cases, it is fine. With AI, identifying a real programmer is just 2-3 topics away.

26 years ago.

tried a cherry and it’s good. Going to do some cherry picking next week

I remember reading about this and goosebumps 🔥

iykyk

i am liking this book.

Tomorrow , it is going to be 3rd week of being sick. Now, I am recovering. I am far better...you can guess from my X activeness. My throat is fine. I am able to sleep longer at nights 4-5hrs. Cough is still there when I try to speak but still way better than before. I am able to hear clearly now (yes, I had blocked ear for 2 weeks). During 2 weeks, I had to skip 1 conference opportunity & 2nd due to visa. Cancelled a personal holiday (all money got wasted). Took work off (which is fine. I am thankful to have supportive team). Hubby had to took off and then wfh to manage me. The house is a mess, rn. Our neighbours were v sweet to offer help/food but my appetite was not there. yesterday, I stepped outside after 2 weeks to see Sun (husband had yesterday calling me vampire as I was just sitting in dark 24x7 for 2 weeks). What all worked: steam, air purifier, humidifier, lemon & honey water, Vicks, candies for cold/chest congestion.

UX designer in my network - which is your goto AI tool for redesign or getting feedback on the designs? I am already using gtp5.5 but not so happy with the results. I also gave agent role. tia

How US gov is going to track? Engineers who worked in it have it running on their local , or access of repo 🤷‍♀️

he comes from my v wealthy background 🙏🏽. He had the appetite to take financial risks Check his parent’s profile.

🙏🏽

One of the best use or example of Agentic AI I can think of is "Incident Triage Agent". As an engineer, when you get page at 2:30 AM for a broken service then most of the time goes in just - understanding the logs, retrieving important information etc. It take atleast 30 mins before engineer starts working on fix. With agentic AI, the research part can be outsource to agent and time can be reduced to 88-95% . Incident Triage Agent will gets a cloudwatch alaram and from there it will get on the work to 1. fires to SNS, 2. invoking Lambda hosting the agent, 3. The agent reasons with Claude (via Amazon Bedrock or any) in a ReAct loop, 4. Claude decides what to check next( no hardcoded runbook ) 5. It reads evidence (metrics, logs, deployments, traces) read-only, it can't touch infra 6/7/8. Posts structured findings to Slack  and pages to pagerduty with all finding The interesting part isn't the LLM call. It's the autonomy: the agent picks which tool to call based on what it just found. A design lesson that I learned: once the agent owns paging, a crashed agent means no page. So I added an independent health-monitoring path CloudWatch alarms on the Lambda's own errors, timeouts, and a dead-letter queue page on-call directly. The agent enriches when healthy; CloudWatch covers it when it's not. Never silent.