product at @openai, ex-@stripe, ex-@ycombinator, Free e-sign https://xitter.cf/t.co/PjxhLnwr5J
Joined October 2010
- Tweets5.9K
- Following801
- Followers11K
- Likes14.3K
This is a great idea. Here’s the exact prompt i used
“Look through all my past conversations and memory. Based on what you know about me, help me generate an image of a bookshelf with 12 books you think describe me as a person/reflect who i am. They don’t have to be books I’ve read or discussed with you, pick any books you think are relevant. Give me a final image of a bookshelf with 12 books in 3 rows, 4 books each row”
If you’re building any sort of persistent or background agent, you’ll understand how important this ship is.
I’m on vacation for a few days and all this sh*t drops
So basically, today PHP got two critical bugs that let anyone run code on your server without even logging in, which is bad, and then the details leaked before the patch was even out, which is worse, and it works on WordPress and MediaWiki and Magento and Joomla and Laravel, and Censys says like 1.5 million servers are running a version with no fix, so the NSA was like "take your system offline," and the FBI was like "take your system offline," and CISA and the UK and CERT-EU were all like "take your system offline," and then the Prime Minister of Japan, the actual Prime Minister, posted at 5:55 in the morning telling 48.8 million people to take their system offline, so everyone took their system offline, like the Dutch cybersecurity agency took its own website offline, and the Dutch parliament went offline, and the Port of Rotterdam went offline, and Finland and France and Denmark started pulling government sites, and Wikipedia went down, and Archive.org went down, and Asana and GOG and Something Awful went down, and the University of Vienna's webmail went down for some reason, and that's just PHP. Because GitLab also dropped a CVSS 10.0, which is the highest score they give out, for a bug that's been there since 2014, so that's every self-hosted GitLab for twelve years, and then LiteLLM 1.107.0 got hijacked and shipped a credential stealer and a remote access trojan in every single wheel and the Docker images, including latest, and you didn't even have to import it, just installing it was enough, and it's the second time this year, and then Railway paused every deployment on the entire platform because someone got into the environment variables, so now everyone has to rotate every secret they've ever stored there, and TiDB Cloud said someone was walking around inside with stolen credentials for basically a month and an extortion group says they took 10.6 terabytes, and Zeabur got breached for the second time in two months and is just shutting down, like the whole company, and Node.js patched remote code execution and a TLS bypass, and CISA says Plex is being actively exploited, and someone took over 300 Tor exit relays to steal SSH passwords, and SafePay ransomware has hit over 600 critical infrastructure orgs, and India's biggest bank filed a data loss incident, and Japan has been in an official state of emergency in cyberspace since October 9, which honestly explains the 5:55 AM post. Oh, and eMudhra, a certificate authority that's issued over 60 million certificates, whose entire job is proving websites are who they say they are, was supposed to be using private keys that are random numbers over a hundred digits long, and one of them was 4. Just 4. And the others were like 66 and 237. Anyway, it's Wednesday.
Fun fact: some of these we just shipped a few hours ago.
Try out OpenAI’s Agents API to see all the fun stuff we shipped!
🗣️ THIS WEEK IN AGENTS API 🗣️
⚡ 6-Astra Ultrafast: now available in the Agents API!
📷 Sandbox snapshots: Restore OpenAI-hosted environments after idle with files intact and keep the work going.
🤑 Better billing: you won't pay for environments that are idle
🔑 Environment keys: Create environment keys programmatically (not just through the dashboard)
📈 Artifact SDK: Create and edit spreadsheets, slides, and Word docs in OpenAI-hosted sandboxes
🔎 Turn-level history: Fetch the items from a specific root-agent turn, including subagent interactions
🔒 Vault filtering: Filter vaults by metadata
Astra is amazing with Browser Use - now developers can integrate in a first class way
What happens if you invest $1,000 in the top performing stock from the last month rather than the S&P 500? 🤔
Apparently you'd be up over $6 million compared to the S&P 500.
I had an agent backtest a strategy of looking at the past month's top performing stock and dumbly investing $1,000 into it. That strategy, since 1998, generated $8.8 million while the S&P 500 generated $2.2 million.
Some caveats:
- there's some stocks the agent couldn't pull/find data for
- we assume that if we can't find the data (e.g. a buyout) that the stock went to $0 to be conservative
- the gains are mainly from a few outliers, like Apple and NVIDIA (if you remove the top outliers, performance gets worse but still beats S&P500)
Built with the Agents API!
I open sourced it, check it out github.com/azianmike/monthly…
and if you want to see the HTML page the agent built: azianmike.github.io/monthly-…
Inspired by xitter.cf/gurgavin/status/210939…
My favorite use case for Astra + Computer Use: help me re-organize my Slack groups + channels.
It'll click around on the UI and group everything.
For example, we launched Bedrock Managed Agents with AWS and there's tons of disparate internal + external channels that Astra helped me organize.
As the number of doctors consulted approaches infinity, the number of foods you’re allowed to eat approaches zero.
💸 I built a stock researcher that costs ~$1 with 1 API call. And then I open sourced it.
I'm so busy now that I don't have time to look at stocks and earning reports. But now with the help of OpenAI's Agents API, I can spin up a long running agent to go search the latest earnings, research the quarterly/annual earnings reports, and build an interactive HTML report.
It runs for over 20 minutes all by itself.
Oh and the Agents API also made the marketing video too.
This but you honestly don't need to use ultra mode with computer use. I regularly use Astra with medium/high thinking and computer use it and it's great.
internal memo:
If you have not yet, please use ChatGPT Astra 6 with computer use on ultra mode if you are doing knowledge work tasks.
This is the first time since Claude Code in Dec 2026 that I've felt a step function change in capabilities - ability to use AI just operates entirely differently.
These are expensive background tasks (many run for 1hr+), but completely fundamentally change ability to use models.
Warning: never ever ever give up your understanding to these models -- if you don't understand what it's doing, or why the outcome is the right one, you are setting yourself up for long term failure.
