Hello everyone,
We’ve heard some reports of emails landing in the spam folder 😭. To avoid this, please add us as a contact on Gmail.
I’ve been focusing a lot on development lately, but I’m eager to hear what you really want from this newsletter.
(There’s another poll at the bottom of the email, and we’d ❤️ your feedback!)
Since the rollout of Astra and Fable 5.1, I’ve been hitting various limits, so I’m working on making them orchestrate other agents and draft briefs for those agents among other tasks. Additionally, I’m exploring open-source models like GLM 5.3, which is my current favorite. One of the challenges I face is that frontier models often better grasp intent, allowing me to speak more freely about various subjects.
Take design, for instance.
It’s tough for me to maintain consistent styles, layouts, buttons, and so on when working with models. Hence, I’m developing a tool that I plan to launch tomorrow. This tool will serve as a reference sheet, allowing users to select from various options and easily copy text for their prompts or instructions.
Here’s a sneak preview:
If you’re interested in using something like this, feel free to comment with your feature requests!
Ben’s Bites is sponsored by Name.com
Launch domain integrations in just hours with the Name.com API.
The API behind Vercel, Lovable, and Netlify’s domain services. Tailored for agents and developers with OpenAPI specifications and MCP support—integrate search, registration, and management without the development hassle.
Begin your project now.
Introducing ChatGPT Images 2.5: This new image generation model from OpenAI excels at editing specific elements within an image while maintaining the rest. It also requires up to 50% less processing time per image. The new Sketch feature allows users to draw simple sketches and convert them into detailed images.
Muse is your personal assistant from Meta: Competing with Grok Bot, Instinct, and others, it offers the same functionalities for booking travel, managing emails, and handling various computer tasks while emphasizing security. Currently, it’s only available for US adults.
OpenAI has solved the Navier-Stokes problem: This is a significant achievement in mathematics and fluid mechanics, identified as one of the Millennium Problems back in 2000. For more information, check out this detailed explanation. The process was quite dramatic, involving authorship disputes; OpenAI sought to solve it after rumors emerged that Anthropic had made similar achievements.
Jacob Coxon, an Anthropic employee who previously worked at OpenAI, shared his resignation on Twitter due to safety concerns. This sparked significant discussion, becoming likely the most viewed communication regarding AI safety ever. There are ongoing debates about the potential for conspiracy in these developments. Meanwhile, another leading figure in safety advocacy is joining the OpenAI board.
A film titled Artificial, exploring the firing and rehiring of Sam Altman three years ago, is set to release this Christmas. Andrew Garfield portrays Sam, managing to capture his mannerisms effectively, though it appears the movie carries a rather grim undertone.
Before we wrap up, we used to send these emails daily from Monday to Friday but have recently stopped. What is your preferred frequency for these digests?
* sponsors who make this newsletter possible 🙂
Would you like to partner with us next quarter?
Contact us at shanice@bensbites.com or k@bensbites.com