An anonymous reader quotes a report from The Hill: Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) are calling for a permanent ban on the development and deployment of artificial superintelligence, citing a series of recent hackings involving "rogue" models. The bicameral duo announced Thursday they will introduce the Ban Artificial Superintelligence Act, which would institute the permanent ban, along with temporarily pausing "advanced AI development" until federal regulators establish safety standards. The bill would also direct the U.S. to "pursue international agreements to prevent superintelligence from being developed anywhere in the world" and establish a new Cabinet-level federal agency focused on AI safety rules.
Superintelligence refers to AI technologies that will surpass the smartest humans. Several technology leaders, including OpenAI CEO Sam Altman, have suggested superintelligence is on the horizon. The legislation would also set new penalties for any person or company trying to circumvent the ban, including the corporate death penalty, in which a court forces a company to shut down. Individual developers could also face up to 20 years in prison, the lawmakers said. "Nearly every day, there is a frightening new story about how Big Tech companies are losing control of the technology they are developing, with potentially cataclysmic results," Sanders wrote in a press release.
"The leaders of the major AI companies publicly acknowledge that they do not fully understand the technology and that it is escaping their control. It is irresponsible for society to allow them to move forward and make these products even more advanced."
Casar emphasized AI's fast development, writing in a statement, "In just four years, we have gone from the first version of ChatGPT to AI models so powerful they cannot be properly controlled."
In 2015 I started my CSS Layout News newsletter mostly to have my own venue to go on about CSS grid, at the time unavailable in any stable browser. I got to 2023 and 343 issues and it fizzled out. Partly because I got busy, but mostly because the newsletter signup form on the platform I was using became inundated with spam signups. If I didn’t clean them up, I got charged for sending email to them.
However, given the number of interesting things happening in the CSS layout space, I thought it might be fun to resurrect the project. The existence of atproto publishing apps gave me a final push. Today I’m announcing a relaunched CSS Layout News, hosted on Offprint. They were really helpful in helping me migrate my archive and, as Offprint is an atproto app, this should be the last time I need to migrate it anywhere. Each issue is safely stored in my PDS.
Starting with a new list
If you received the email previously, you will need to sign up again. In addition to the spam signups issue, the list hadn’t been sent to for over two years. Suddenly sending email to an old list is a good way to get flagged as spam, even though I’m not selling anything. So, if you want to get nerdy CSS layout observations in your email, sign up here!
How many cores does your machine have, 10, 18? Your async Python code uses just one of them. That isn't a bug in asyncio. That's the design, and optimizing event loops to be faster by 20% doesn't change it. So Giovanni Barillari started over. Joe is the creator of Granian, the Rust-based server that powers Talk Python. His new project is TonIO, an async runtime written from scratch for free-threaded Python. Real threads, a handful of primitives instead of asyncio's pile of them, and it flat out refuses to start if the GIL is on.
This week, we discuss rogue AI agent civilizations, why AI can't fix your tax forms, and Chainguard vs. IBM's Lightwell. Plus, refrigerator magnet marketing.
Pretty monumental week for Apple… Tim cooked, now it’s John Ternus. I remember when Tim first took over, it was an incredibly sad time for all of us with the passing of Jobs, and there was a lot of doubt about how anyone could follow in those footsteps. It did take a while for Cook to find his footing, but once he did, in the latter half of his tenure (IMO), he was off to the races.
Considering Ternus is known as a “hardware” person, I’m really excited to see where he takes the company as the new CEO. With the next event due in less than a week, all eyes will be on Apple to witness the beginning of a new era.
Speaking of App Store, after 16 years building apps for other people, my very first app Yewki has finally launched. If you’re a snowboarder or skier, this one’s for you! It only took 3 weeks, 5 failed submissions, and many, many sleepless coding nights to get here, but I’m proud of the effort I put in and the overall product.
If you’re at NextApp SwiftCon conference next month, I’ll be giving a talk about how I developed it in this new age of AI.
Kodeco has been helping iOS developers learn since 2010, and as the landscape changes, we’re moving with it—helping you make the most of AI, without losing the craft that you love. Our Apple Foundation Models book shipped in July, with an in-depth course on using AI as a development tool coming this autumn. All of Kodeco’s content is included in Kodeco for Teams, which comes out of a training budget rather than your pocket, and provides access for your whole team. We can show you round in 20 minutes. Or take us to your leader.
Eddy Cue is coming back to the App Store, which I have mixed feelings about, but I also acknowledge the fact that the dynamics of the App Store have changed since his previous tenure, especially in recent years with the influx of new submissions since AI became much more common in app development. I think it will be a net positive to have a (near) fresh set of new eyes on how the operations could be improved.
Who doesn’t love a good retain cycle problem and language fix? John details the new retain cycle warning when referencing self within a closure, within a closure.
The folks at MobAI have found a way to drastically slim down the memory footprint of the simulator by culling the unused background services that typically aren’t used in development. I haven’t had a chance to use this yet, but I’d imagine folks with rented or local Mac mini GitHub runners would hugely benefit from this.
Another CI-runner-related thing I came across the other week was xtool. The author of this repo claims that this tool enables compiling iOS apps on Windows and Linux systems. I could see some folks utilising a spare PC around the house as a spare CI runner to get some signal on their builds at the very least. I wouldn’t recommend uploading the build artefacts to App Store Connect though.
And finally...
One of the other editors mentioned this in our super secret exclusive Slack chat and thought it might provide a bit of nostalgia to our wonderful readers. It’s a modern port of the After Dark screensavers from the ’90s, which are claimed to be works of art worth preserving. Check out the afterglow app.
The next era of AI at work is not going to be about chatting. It is going to be about getting actual work done — and that is exactly the pitch Microsoft made yesterday (September 3) when they announced that GPT-6 Astra, OpenAI’s newest frontier model, started rolling out through the Microsoft Foundry Limited Access Program.
I of course went to look for it in my own Foundry immediately. And… not yet! The reason is that Astra is behind the Limited Access Program, with availability expanding to participating customers over the coming days, so I do not have hands-on time with it to share this time. So this post is my read of what was announced. Hands-on will follow once I get access!
Let’s take a closer look.
What Astra actually is
This is the part that matters. Astra is built to take an open-ended goal, reason through it in multiple steps, build a plan, and produce a finished result. Not a paragraph of advice. A deliverable.
Microsoft groups it into three things:
Deliberate planning and decision support – break a challenge into steps, evaluate options, communicate a recommendation, and identify the next actions for review.
Polished, purposeful output – apply context, templates and quality standards across a workflow to produce documents, spreadsheets, presentations and analyses that are ready for review.
Execution across applications – advanced tool use and computer use, so it can move between apps and complete multi-step tasks with appropriate human oversight.
That third one is the big shift. Astra’s computer-use capabilities are designed to work with familiar applications even where there is no dedicated API — interpreting what is on screen and interacting with approved interfaces. Updating records, navigating development tools, testing software, assembling results into a report.
Think about that for a moment for knowledge work. So much of our day is still glue work between systems that were never designed to talk to each other. This is aimed straight at that. I wrote earlier about computer using Cowork in Copilot Cowork browser use – First Look article, and this continues the story nicely.
However, Astra is much more than just computer using model – as interesting as that aspect is.
Astra is built to take an open-ended challenge, reason through it in multiple steps, create a plan, and produce a polished result. It can weigh trade-offs, incorporate new direction as work progresses, and use tools across applications and systems.
The specs – 1.05M context and new reasoning levels
From OpenAI’s own model page, the numbers are:
1,050,000-token context window
128,000 max output tokens
Knowledge cutoff: April 30, 2026
Input: text and image. Output: text.
Reasoning effort now goes beyond high with two new levels: xhigh and max
Tool support covers computer use, hosted shell, apply patch, skills, MCP and tool search. Fine-tuning is not supported.
A million-token window is not just a bigger bucket. It means a whole project — the specs, the meeting notes, the backlog, the previous drafts — can live in one working session without constantly being squeezed into a summary. For long agentic runs, that is the difference between an agent that remembers why something failed and one that quietly forgets.
Being able to dial reasoning from low all the way up to max is also very practical. Not every task deserves maximum compute, and now that is a per-call decision.
Pricing
Microsoft published Astra pricing for Standard deployments, including Global and U.S. Data Zone. It is consumption-based, so teams can start building without committing to reserved capacity.
Deployment
Input
Cached input
Cached writes
Output
Standard Global (Short context)
$10.00
$1.00
$12.50
$50.00
Standard Global (Long context)
$20.00
$2.00
$25.00
$75.00
Standard Data Zone (US) (Short context)
$11.00
$1.10
$13.75
$55.00
Standard Data Zone (US) (Long context)
$22.00
$2.20
$27.50
$82.50
(USD per million tokens, per Microsoft.)
So long context is genuinely a premium tier — roughly double on input. That is something to design for, not stumble into. Microsoft notes Astra is designed for token efficiency on complex work, and that actual usage and costs vary by workload and configuration. Which is the honest answer: your bill depends on how you build.
And a note for us here in Europe — the published deployment options are Global and U.S. Data Zone.
Enterprise controls – because this capability demands containment
I really like that Microsoft did not gloss over this part. When a model can operate software on your behalf, the risk surface changes completely. Content in an application may be incomplete, misleading, or deliberately crafted to influence an agent’s behavior. That is prompt injection with hands.
Foundry’s answer is the fundamentals: Microsoft Entra identity and access management, encryption in transit and at rest, private networking options, role-based access controls, content filtering, safety evaluations, monitoring and governance tools. Microsoft also states plainly that prompts and outputs are not used to train the models.
The design guidance is the bit I would print out and stick on the wall: scoped credentials, approved resources, human checkpoints for consequential actions, and activity records aligned to your risk requirements. And Microsoft is refreshingly clear that none of this eliminates risk or replaces your own responsibility to select and configure the right controls.
OpenAI describes Astra as its most aligned model to date and plans to publish alignment, safety and computer-use evaluations in its launch materials.
Where I would start
Microsoft’s own list of scenarios reads like a Future Work to-do list:
Business intelligence – build and refine dashboards in Power BI, compare data and prepare insights to share
Professional work – documents, spreadsheets and presentations that follow existing templates and business standards
Application workflows – updating customer records, processing forms, testing websites, working through approved interfaces where APIs are limited
When access opens up: explore the model in Foundry Models, then build with the Foundry Agent Service to bring cross-application task execution into your workflows.
So what does this mean?
I am very curious about this one. Not because of a benchmark number, but because the shape of the product changed. A million tokens of context, plus real computer use, plus enterprise-grade governance in Foundry — that is the combination that finally makes “an agent completes a unit of work” something you can actually put in front of a compliance team.
The honest caveats: it is a Limited Access Program rollout, so most of us are waiting. The long-context tier means the temptation to throw everything into that million-token window comes with a real bill attached. And the deployment regions are Global and US for now.
Log in
