Building AI systems for real businesses. Shipping products in public. https://xitter.cf/t.co/9WX0H4XuFB | https://xitter.cf/t.co/22Yua9rkfi | https://xitter.cf/t.co/q3qfGKdR8D

Earth
Joined June 2021
Agent wars is about to get very real.
Maybe we should start a competitor to Amazon 🤔
11
This insider will be on mainstream media by Thursday telling us how AI is going to kill us all.
We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work. At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated. We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.
14
Ya that seems like a really basic use case for the horse power you get with Hermes. It’s currently running my software factory 🤣
Friendly reminder that Hermes was never built or intended to be exclusively for the "personal assistant" agent that can just read your emails and nothing else. I fully intend and have plainly stated many times that I always built hermes to be the most powerful AI Agent. "Normie" consumers can leverage that and we do what we can to make it an experience they can approach (the mobile app will be almost exclusively focused on consumer) - but they are not our only demographic. I want the scientists, the cybersecurity engineers, the developers, the knowledge workers, the creatives, and everyone else to be able to have an agent that they can leverage that is maximally useful to them too. Not just your mom. (She called btw you need to make your bed)
1
31
This just looks like theatre at this point.
🚨 NOW: Anthropic chief DARIO just walked up and said President Trump helped put his Super Intelligence worries at ease "If we do this right, we work with the President and everyone here, we can win safely!" Imagine that, all it took was DONALD J. TRUMP getting involved and the tables are turning! No more SI dooming, no more surrendering to China. America will remain number one BOTH for national and economic security, as long as Trump is president 👏🏻
37
OpenAI: your agent can keep working after you close the laptop. Me: perfect. Also me, 11 minutes later: opens the laptop to check whether the agent is still working. xitter.cf/OpenAIDevs/status/2104…
You can finally close your laptop now and your agents will keep working. Codex cloud environments are here. Reusable environments mean less setup and faster starts, with your repo, dependencies, scripts, and settings already in place.
33
OpenAI launched Dots, Codex Cloud, Space, Pages, Ultrafast and 20+ updates in one day. My current strategy is to open six tabs, understand none of them and confidently tell myself I’ll test everything tomorrow.
44
Putting a full app inside ChatGPT changes the first step. People can start with a problem, then use the tool without leaving the conversation. For small business software, that may matter more than adding another AI button to the old app. xitter.cf/thsottiaux/status/2104…
We are opening up our platform and you can now build full native apps with plugin extensions, and ship them right in ChatGPT. We have over 1.2B weekly users and will surface relevant plugins right in the conversations.
1
34
OpenAI after 20+ DevDay announcements: here are 10 side quests and five Codex Micros. Apparently the keynote was the warm-up. xitter.cf/OpenAIDevs/status/2105…
Ready for a DevDay side quest? 10 builder challenges. 5 Codex Micros up for grabs. Complete any 5 side quests to earn 1 raffle ticket for each. 48 hours. Join from anywhere 🌎
1
64
OpenAI cutting the $200 Pro plan from 20x to 10x usage without cutting the price matters beyond OpenAI. Once one major provider normalizes shrinking the allowance mid-plan, every other AI subscription has cover to do the same. Builders need usage we can actually budget around.
1
15
ChatGPT Space might be the sleeper launch today. It gives people, ChatGPT and always-on agents a shared place to work from the same knowledge instead of passing screenshots and copied notes around. That could clean up a lot of messy project handoffs.
1
22
I’ve been running agents across four machines. Parallel work is powerful, but it gets messy fast when you can’t see what is running, waiting or blocked. The new Codex /agents view is aimed at exactly that problem. This one feels practical. xitter.cf/OpenAIDevs/status/2104…
Codex CLI just got a major refresh, with a new look and powerful new capabilities. We’re continuing to invest in Codex CLI for developers who live in the command line. New full-screen interface, better readability, and ways to manage parallel work, all inside your terminal.
29
300 tokens per second is wild. I’m still more interested in what the task costs from start to finish. Speed helps. But for long-running work, better caching, fewer retries and choosing the right model may save more than raw output speed. xitter.cf/OpenAI/status/21049939…
This is Ultrafast. Our premium speed tier, Ultrafast offers up to 8x faster token generation (300 tokens per second) in Codex and up to 6x in the API.
1
21
This is the one I want to test first. The hard part of overnight agent work isn’t starting the task. It’s preserving the environment, evidence and handoff so you can trust what happened when you come back. xitter.cf/OpenAIDevs/status/2104…
You can finally close your laptop now and your agents will keep working. Codex cloud environments are here. Reusable environments mean less setup and faster starts, with your repo, dependencies, scripts, and settings already in place.
24
One of the quieter DevDay launches could matter a lot: Sign in with ChatGPT. OpenAI says Plus and Pro users can use their plan allowance inside 16 partner tools. If this works well, paying for the same intelligence over and over in separate apps may finally start to fade.
29
OpenAI just announced a Decisions API. It sounds a lot like Jev from @TypeSafeAI: give AI context, define allowed answers, then let your code decide what happens next. That’s a very different job from asking a model to write. The decision layer is becoming its own category.
1
123
OpenAI just launched Dots: always-on agents with their own cloud computer and access to 4,000+ apps. This is close to what I’ve been testing with our own agent fleet. The part I care about most: clear limits on what it can do, when it asks first and what it never touches.
22
OpenAI DevDay starts at 1 PM ET today. OpenAI says we’ll see what’s next in ChatGPT, Codex and building tools. I’m watching for what helps small teams get useful work done with less setup and less babysitting. I’ll test the launches against what we’re building.
52
I am less interested in AI replacing people than in AI helping people do more. Good systems can give a small team better research, support and follow-through without hiring a large technical team. That feels like a better use of AI.
2
32
Small businesses do not need ten AI tools that never talk to each other. They need a few simple systems that share the right context, keep a clear record and know when a person needs to step in.
1
3
18
Building AI systems for @houseofgarda has taught me this: The hard part is not making AI do a task. It is fitting that task into how people already work. If the new system adds more work, it is not helping.
1
1
41