Likes to talk AI & automation 🇺🇸🇸🇬🇲🇽 + other places
Mexico City
Joined August 2009
- Tweets905
- Following5.6K
- Followers617
- Likes10.5K
Pinned Tweet
Most people aren't missing a perfect AI tool. They're missing the system around it. I'm launching System Addict, a weekly newsletter teaching you how to build workflows with context, handoffs, evals, scheduled tasks, and human judgment. Sign up: systemaddict.beehiiv.com/sub…
Jamaur Bronner retweeted
Dario turned out to be right. Maybe even a little cautious. But look at the comments. All so very convinced that they knew more and better, and what an idiot he was.
Inside the usage tab in the Claude app, it gives you an estimate of when you're likely to run out of your quota. Super helpful for pacing yourself.
Jamaur Bronner retweeted
We will give one banked reset for every day you don't have access to Astra on your paid ChatGPT plan, starting today. Team is moving mountains to give access as fast as we can.
First one will land in ~ 3 hours. There is still time to create your account if you don't have one.
Decided to turn this benchmark chart into a score differential, since it's easier to evaluate the deltas between the leaders of each benchmark. 3.8 Flash is certainly impressive for a flash model, though it - and the other frontier llms - lag Opus 5 on a few popular benchmarks.
🤖 Made with AI
Today we’re introducing a new 3.8 Flash model, our 3rd Flash release in just 6 wks.
It delivers significant leaps from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning. On DeepSWE v1.1, it outperforms most larger frontier models in autonomously solving complex engineering problems end to end, at a fraction of the cost.
More details here: blog.google/innovation-and-a…
Felt this problem too. I ended up working with Codex to build a tool called Mosaic so my agents can search Gmail, Drive, and Calendar across all my Google accounts (personal and work).
🤖 Made with AI
I believe all knowledge work is, or is turning into, systems design. It's what I write about in my weekly newsletter System Addict
Cloudflare has some of the best tools for agentic engineering. It's actually crazy how many products they have, and how much is available on their free tier
Introducing Kitesurf: a browser built for agents, running entirely on Cloudflare Workers.
Chromium is too heavy to hand every agent one. Kitesurf is written in Rust, uses 3-7x less CPU and memory, and spins up per request.
Free in beta: blog.cloudflare.com/kitesurf…
You might hate AI.
There are plenty of things about its growing ubiquity that might annoy you. But I dare you to watch this video without tearing up.
AI as an accessibility tool is huge for humanity.
Congrats to the Bland team. I love seeing AI used for good.
Today, we’re launching Bland Speech v3 - The world's first Human Speech Engine.
In @designarena’s Audio Realism benchmark, Speech v3 is the top model, outranking Elevenlabs, Grok, Cartesia, and OpenAI.
Trained on over 100 million real, human conversations: businesses, developers, and creators can now create any audio they can imagine.
To showcase this, we used 5 seconds of old footage to help James, a 49 year old father who recently suffered a stroke, get his voice back.
Try it for free at bland.ai/speech
Full length documentary below🧵
I just reserved my Cloudflare Wallet tag: jamaur.cloudflare.pay.
Reserve yours now at cloudflare.pay
AI too cheap to meter will awaken a new wave of entrepreneurial thinking. Not just in a money making sort of way. Think of all the little problems or annoyances you have that you can just point AI at to help solve for you (and others) in minutes.
Jamaur Bronner retweeted
DeepSeek silently updated their changelog with a new V4-Flash upgrade 1 hour ago.
Their new Terminal-Bench score is 82.7, a massive +25.8 point leap from its initial April preview score of 56.9.
Currently only available via their API, open weights release will follow shortly.
My chill music player that's hosted on a @ChatGPT Site now has night mode. Check it out!
stillpoint-sessions.jambronn…
Jamaur Bronner retweeted
Want to start a company? Great. Get an appetite for glass. You’re gonna chew it every day.
Seems like there's a step change incoming. This year, agents have become reliable for real work. Token efficiency gains mean more types of work can be done. Pair that with voice AI + faster inference (e.g. OpenAI + Cerebras), and we have the base for entirely new ways of working.
openAI says they’re optimizing GPT 5.6 Sol so every prompt uses less usage, while temporarily removing the 5h limit…
seems like we’re entering a different phase of the AI race…
for the last 2 years it was all about shipping the smartest model and now it feels like the real competition is making those frontier models cheap enough that people can actually use them all day…
honestly, i think that’s a much bigger advantage than squeezing out another few benchmark points..
I think it's great that OpenAI has given out so many token resets (3 in a 24 hour timespan this week), but it's crazy that they've had to defend how they choose to award them. Subs are already incredibly subsidized relative to their value. Banked *and* instant resets are a gift.
We applied the resets automatically because of two reasons.
1. It is not possible to use a banked reset from the web or mobile ChatGPT. We are launching it next week and it will require a bit of education for all our new ChatGPT Work users.
2. The system is under a lot of pressure and full resets are more predictable in how they affect load on the infrastructure. This helps us keeping things stable in the first 48 hours of the launch.
And we will use banked resets once we are back in the normal operating cadence!
I didn't appreciate this in the announcement at first, but Chat now embedded in the app is great for devs. I often do deep research on features with Pro, and then work the insights into what I'm building. Now I don't have to switch between the Codex and ChatGPT apps to do so.
Today we’re introducing the brand-new ChatGPT desktop app. If it looks familiar, that’s because it grew out of the Codex app.
We were deliberate about how we brought the pieces of this app together. Rather than splitting it into separate, disconnected worlds, we designed Work and Codex to share the same powerful capabilities and task history. The difference is in the interface: Codex is built to show you code, diffs, and PRs, while Work keeps those details abstracted away. Pick one, or move freely between them.
When you want to quickly explore an idea or get an instant search result, Chat is now in the app too—with instant and pro models, search widgets, and a seamless handoff into Work or Codex tasks.
Actually, looks like only web search is an option, not deep research. Any plans to add deep research to the desktop app @ajambrosino?