
The acronym ICE for "Integrated Creativity Environment" seemed like a good idea until...

Founder and CEO @reve. Building human/agent creative collaboration at itr8.ai. Former VP of Product at Stability AI. Ex-Adobe. Writer repped by Gersh.

The acronym ICE for "Integrated Creativity Environment" seemed like a good idea until...



GPT Image 2.6 is here. We went through a wishlist of improvements, like speed, editing precision, less AI style, visual annotations, and GIFs. With this, the Pareto frontier is just GPT Image!
BREAKING: GPT Image 2.6 Sunburst by @OpenAI is ranked #1 on Image Arena, overtaking its predecessor Image 2.5 Sunburst. GPT Image 2.6 Sunburst & Flare have also redefined the Preference vs Speed Pareto frontier for both Image Generation and Image Editing capabilities. There is currently no model that is both better and faster than the GPT Image 2.6 models on both frontiers. Key Data Points: - #1 in the Portraits, Architecture, & Typography (GPT Image 2.6 Flare) - #1 in Product & Landscape (GPT Image 2.6 Sunburst) - 2.6 Flare is 2.6× faster than its predecessor for both Image and Image Editing - 2.6 Sunburst is 1.15× faster than its predecessor for both Image and Image Editing Congrats to the @OpenAI team for establishing a new SOTA on Preference vs Speed!

I guess I shouldn't start my ChatGPT conversations with "Hey, can you hear me?"

Today, we’re sharing the training infrastructure behind Reve models. In the past two years, we built and scaled our platform to support frontier-level training efforts. We're releasing this technical report as a helpful reference for other small teams building at scale. reve.com/training-infra

I woke up this morning thinking, "I had the craziest dream about flying cars." How will any future product announcement ever hope to compete?

I just used itr8.ai to compare Jev with @OpenAI's new Decisions API. The upshot: • Jev is slightly faster, but unless your use case is extremely latency-sensitive, not meaningfully. • The Decisions API is slightly more accurate. (Jev just can't do date math.) • If you need to classify images, the Decisions API is the clear winner. No comparison. Overall, if you're already using OpenAI APIs, I think it's pretty hard to make the case for Jev. Anyone finding different results?

I'm looking for a kick-ass product person who wants to do two things: 1. Get deep into the operations side of early-stage startups. 2. Completely rethink HCI. You don't have to be a traditional PM, but you do have to be: • Product-obsessed. • Design- and/or engineering-adjacent. • AI-pilled (though it's OK to hate the term "AI-pilled"). • In person every day in Palo Alto. (We have a beautiful office, great food, and an incredible, talent-dense culture.) Reve's next chapter is going to be big, and we're moving fast. DM me to learn more. (No contractors, please.)

I'm surprised the new Siri is getting so many positive reviews. I'm pretty disappointed. Maybe it's better than the old Siri (I wouldn't know; I seldom used it), but it's significantly worse than all the other agents in my life. Am I missing something?

I love building games with itr8.ai. I can have all my assets (including music) and annotations on canvas at once. I'm about two weeks away from launching what I think will be the most powerful and creative agent harness in the world.

Four flights in four days. I could not be a bigger fan of @Starlink.


I consistently see the industry undervalue agent harnesses. This is where a lot of the innovation will shift. @nvidia's AVO harness demonstrates exactly this dynamic by getting 100% on ARC-AGI-3. developer.nvidia.com/blog/nv…

Scout or Rivian R2? Both are reserved, but only one will get delivered.

Over the weekend, itr8.ai got Gemini Flash 3.7 as well as Qwen 3.8 Max and DeepSeek V4 Pro (to go with GLM 5.2 and Kimi K3). But most interesting was the fact that itr8 added these new models to itself. The second screenshot is a browser instances controlled by GPT 5.6 Sol inside itr8. To test Gemini and DeepSeek at the same time, it created a project moderated by Gemini 3.7 and asked it to verify DeepSeek V4 Pro. itr8 is to the interaction layer what @OpenRouter is to the server.