Discover the Best AI Tools

We test the top AI tools for writing, video, images & music — so you don't have to.

Explore AI Tools

ARC Prize Just Fact-Checked OpenAI's 99.9% — GPT-6 Astra Really Scored 62.7%

Image
  When OpenAI launched GPT-6 Astra on September 3, 2026, one number stole the show: 99.9% on ARC-AGI-3, the benchmark designed to be nearly impossible to game. Nvidia CEO Jensen Huang cited it days later when he declared on X that "AGI has arrived." Now the people who actually run that test have checked OpenAI's homework — and the number came back very different. What ARC Prize actually found The ARC Prize Foundation independently reran Astra on ARC-AGI-3 and scored it at 62.7% on its standard, provider-neutral harness — the same setup every other lab's model must clear. On OpenAI's own provider-adapter harness, the same model scored 98.6% at maximum effort and 99.9% at high effort. Same model, same effort, different scaffolding: 36 to 45 points of difference. ARC Prize's verdict was blunt: the 99.9% figure is not a comparable score, because it was not produced under the conditions every other model faces. The foundation published both runs. Why the harness ch...

Walmart Just Opened Its Store to AI Shopping Agents — While Amazon Shuts Them Out



Walmart logo




 Walmart is betting that your next shopping trip may not involve you at all — at least not directly. On October 21, 2026, Walmart CEO John Furner announced at the Masters of Scale Summit in San Francisco that the retailer is opening its checkout to approved AI assistants. The move sets up a sharp contrast with Amazon, which has been moving in the opposite direction.


What Walmart launched


The centerpiece is the Walmart Agent Gateway, an MCP-based server that lets approved AI assistants search Walmart and Sam's Club inventory and complete purchases. MCP is the open standard Anthropic created, and Walmart is starting the Gateway as a beta for companies and developers. Walmart already works with ChatGPT, Gemini and Meta's Muse, as well as its own Sparky assistant, so shoppers may soon ask a familiar assistant to handle a routine order instead of opening an app and searching aisle by aisle.


How Agent Profiles keep shoppers in control


The second piece is Agent Profiles, rolling out next month in November. Each AI agent gets its own profile inside the shopper's Walmart account, and purchases made by that agent are shown separately from the shopper's own orders. That separation matters: it makes it easier to see what an agent bought, when, and why. Shoppers can also require approval for each transaction or set a spending limit, so an assistant cannot quietly run up a large bill. Walmart says agents never receive payment-card data, only get access to what the task needs, and agents that break the rules are blocked. The system works for any Walmart account, with no Walmart+ membership required.


Amazon's opposite bet


Amazon, by contrast, has been blocking outside AI agents from buying on its site, including Meta's Muse. That is a very different theory of shopping. Where Walmart is building a doorway for approved assistants, Amazon is keeping the door closed and steering shoppers to its own storefront and tools. For readers following this split, Amazon's earlier move against outside agents is a useful backdrop, and it shows why Walmart's announcement is not just a feature update — it is a strategic choice about who gets to sell through an assistant.


Why it matters


If this model works, shopping could shift from browsing and checkout pages to conversations with an assistant that already knows your preferences. That could save time on repeat purchases, but it also raises the stakes for control, transparency and trust. Walmart's answer is to keep the shopper in the loop with separate agent profiles, approval options and spending limits, rather than handing an assistant an open wallet. How strict those controls feel in practice, and how many developers join the beta, will decide whether agent shopping stays a novelty or becomes a normal way to buy groceries and household goods.


Short take


Walmart is inviting AI agents in, with rules and limits attached. Amazon is keeping them out. The next test is simple: which approach shoppers trust more when an assistant, not a person, clicks buy.


Comments

Popular posts from this blog

AI Tools Every Accountant Should Use in 2026

Top 5 Best AI Video Generation Tools in 2026

7 Best AI Website Builders in 2026