AI Agents Under Review, But the Corporate World Leans In

Ai5ers,

This week in Ai5: Gartner busts the myth of intelligent AI agents, CMU launches a benchmark to test them, and Amazon hits a milestone with one million robots. Meanwhile, Apple and Meta escalate the AI talent war, the Senate rolls back federal limits on state AI laws, and Cloudflare gives websites a way to block AI bots. Dive into the signal beneath the noise.

Keep reading!

Today's Ai5:

🧐 AI agents get office tasks wrong, a lot of them aren't AI
💰 Apple and Meta are proving it: AI is going corporate
🦾 Amazon deploys its 1 millionth robot, releases AI model
🤨 US Senate overwelmingly removes 'AI moratorium'
🚫 Millions of websites to get 'game-changing' AI bot blocker

Prompt of the Day 🎨

Celebration of Independence Day:

Generate an image to celebrate the 4th of July. The image should be photo realistic and have three columns: Column 1 should be of a golfer ripping off a tee shot. Have the ball be followed by a tracer in the air and have the ball explode like a beautiful firework. Column 2 should be of a person wakeboarding behind a water ski boat. Have the person spraying water from the wakeboard. Have the spraying water resemble a beautiful firework. Column 3 should be of a beautiful little girl about 4 years of age wearing a 4th of July themed dress playing with a sparkly sparkler. The little girl should be surrounded by older kids about 10 years of age watching the little girl's activity with excitement and the glow of the sparkler shining on all faces.

AI agents get office tasks wrong, a lot of them aren't AI 🧐

According to Gardner, many agents are fiction without the science. "Many vendors are contributing to the hype by engaging in 'agent washing' – the rebranding of existing products, such as AI assistants, robotic process automation (RPA) and chatbots, without substantial agentic capabilities," the firm says. "Gartner estimates only about 130 of the thousands of agentic AI vendors are real."

For a reality check, CMU researchers have developed a benchmark to evaluate how AI agents perform when given common knowledge work tasks like browsing the web, writing code, running applications, and communicating with coworkers.

They call it TheAgentCompany. It's a simulation environment designed to mimic a small software firm and its business operations. They did so to help clarify the debate between AI believers who argue that the majority of human labor can be automated and AI skeptics who see such claims as part of a gigantic AI grift.

The gap between these two positions is due to the lack of a way to test how agents handle common workplace activities. Hence the need for a benchmark, which suggests AI agents have a way to go before they're truly useful.

Full article by Thomas Clatburn

Apple and Meta are proving: AI is going corporate 💰

This week, Bloomberg reported that Apple is considering using AI technology from outside firms to power an updated version of its assistant, possibly sidelining its own in-house efforts, as part of a broader push to keep pace in the AI arms race.

The report comes on the heels of Meta CEO Mark Zuckerberg's aggressive recruitment drive to poach top AI researchers and engineers at rival shops. He has offered as much as $100 million to pull outside talent into the Meta fold. On top of that, Zuckerberg had reportedly tried to acquire the AI startup Safe Superintelligence, but after the offer was turned down pursued its CEO instead. Meta also recently invested $14 billion in the data-labeling startup Scale AI and hired its top executive, Alexandr Wang, to help lead a new superintelligence unit inside the company.

The intertwining developments capture something important about the business of AI development. They get at how the public views AI's leading upstarts and the tech platforms competing with and courting them. Call it AI's corporate phase.

Full Article by Hamza Shaban

Amazon depoys its 1 millionth robot, releases AI model 🦾

After 13 years of deploying robots into its warehouses, Amazon reached a new milestone.

The tech behemoth now has 1 million robots in its warehouses, the company announced Monday. This one millionth robot was recently delivered to an Amazon fulfillment facility in Japan.

The company also announced it’s releasing a new generative AI model called DeepFleet for its warehouse robots. This AI model, which can coordinate the robots’ routes within the company’s warehouses more efficiently, will help increase the speed of its robotic fleet by 10%, according to Amazon.

Full article by Rebecca Szkutak

US Senate overwelmingly removes 'AI moratorium' 🤨

U.S. senators voted overwhelmingly on Tuesday to remove a controversial 10-year ban on states’ abilities to regulate AI from the Trump administration’s “Big Beautiful Bill,” reports Axios.

The provision to the reconciliation bill was introduced by Sen. Ted Cruz (R-TX). Many prominent Silicon Valley executives — including OpenAI’s Sam Altman, Anduril’s Palmer Luckey, and a16z’s Marc Andreessen — were in favor of the so-called “AI moratorium,” which they said would prevent states from forming an unworkable patchwork of regulation that could stifle AI innovation.

Opposition to the provision became a bipartisan issue, as most Democrats and many Republicans warned that the ban on state regulation would harm consumers, and let powerful AI companies operate with little oversight. Critics also objected to Cruz’s plan to tie compliance with federal broadband funding.

Full article by Rebecca Bellan

Millions of websites to get 'game changing' AI bot blocker 🚫

Millions of websites - including Sky News, The Associated Press and Buzzfeed - will now be able to block artificial intelligence (AI) bots from accessing their content without permission.

The new system is being rolled out by internet infrastructure firm, Cloudflare, which hosts around a fifth of the internet.

Eventually, sites will be able to ask for payment from AI firms in return for having their content scraped.

Many prominent writers, artists, musicians and actors have accused AI firms of training systems on their work without permission or payment.

Full article by Chris Vallance