
Amazon's AGI Lab Reveals Its First Work: Advanced AI Agents
Mar 31, 2025 9:00 AM Led by a former OpenAI executive, Amazon's AI lab focuses on the decision-making capabilities of next generation of software agents—and borrows insights from physical robots. The Amazon logo on the façade of one of their global headquarters. Photograph:Amazon is still seen as a bit of a laggard in the race to develop advanced artificial intelligence, but it has quietly created a lab that is now setting records when it comes to AI performance. Amazon's AGI SF Lab, which is located in San Francisco and dedicated to building artificial general intelligence, or AI that surpasses the capabilities of humans, revealed the first fruits of its work today: A new AI model capable of powering some of the most advanced AI agents available anywhere.
The new model, called Amazon Nova Act, outperforms ones from OpenAI and Anthropic on several benchmarks designed to gauge the intelligence and aptitude of AI agents, Amazon says. On the benchmarks GroundUI Web and ScreenSpot, Amazon Nova Act performs better than Claude 3.7 Sonnet and OpenAI Computer Use Agent. A major part of Amazon's plan to compete in the AI market is to focus on building agents, and the new model's abilities reflect its efforts to build a generation of tools that can measure up to the very best available.
'I believe that the basic atomic unit of computing in the future is going to be a call to a giant [AI] agent,' says David Luan, who leads Amazon's AGI SF Lab. He was previously a vice president of engineering at OpenAI and later cofounded Adept, a startup that pioneered work on AI agents, before joining Amazon in 2024 when the ecommerce giant took a stake in the company.
Most of the leading AI labs are now focused on building increasingly capable AI agents. Getting AI to master independent actions, as well as conversation, promises to make the technology more useful and valuable. The shift from chat to action is still very much a work in progress, however.
In the past six months, OpenAI, Anthropic, Google, and others have demonstrated web-browsing agents that take actions in response to a prompt. But for the most part, these agents are still unreliable, and they can easily be tripped up by open-ended requests.
Luan says that Amazon's goal is building AI agents that are dependable rather than flashy. The thing holding agents back is not the need for 'more cool demos of interesting capabilities that work 60 percent of the time, it's the Waymo problem,' he says, referring to how self-driving cars needed to be trained to deal with unusual edge cases before they could take to the streets unsupervised.
Many so-called agents are built by combining large language models with multiple human-written rules that are designed to prevent them from veering off course, but also makes their behavior brittle. Amazon Nova Act is a version of the company's most powerful homegrown model Amazon Nova that has received additional training to help it make decisions about what actions to take and at what time. In general, Luan says, AI models struggle to decide when they should intervene in a task.
To improve Nova's agential abilities, Amazon is using reinforcement learning, a method that has helped other AI models better simulate reasoning.
Amazon is also taking inspiration from physical robots with its new models. Laun's team is working with another group at Amazon based in San Francisco led by Pieter Abbeel, a professor at University of California, Berkeley who works on finding AI applications for robotics. Abbeel, a fellow early OpenAI employee, joined Amazon in August 2024 after it invested in his startup, Covariant. Amazon is well positioned to make progress in robotics given the vast numbers of those already deployed in its fulfillment centers.
The release of Amazon Nova Act suggests that Amazon could emerge as a dark horse in the race to create useful software agents. The company was slow in responding to ChatGPT, but it has more recently shown signs of getting its act together. In February, the company announced a new version of its voice assistant Alexa with improved conversational abilities as well as the capacity to automate certain web tasks.
One use case Amazon cited is Alexa helping to book a repair service for a broken oven. Luan says Alexa's new agentic capabilities were developed by his team.
As WIRED revealed last October, Amazon has also done research on how agents might eventually improve ecommerce by automating the process of finding and buying things. Such an agent might preemptively add items to a user's cart based on their interests and habits, Amazon's engineers said.
Besides unveiling the new model, Amazon today announced a software development kit (SDK) designed to make it easier for computer engineers to use Amazon Nova Act to build software agents. The SDK lets developers give their agent specific instructions to help them navigate an internet built for human users. For example, an agent can be instructed 'don't accept the insurance upsell' when booking a rental car.
Ultimately, Luan says, Amazon's agents should become smart enough not to fall for the upsell on their own. 'Nova Act is really like the very first step in that vision,' he says.
Hashtags

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles


Business Insider
an hour ago
- Business Insider
Oracle Stock (ORCL) Sustains Bullish Posture and Charges into the AI Era
Multinational technology magnate Oracle (ORCL) is storming the AI frontier, riding a massive wave of demand for its cutting-edge cloud infrastructure and AI-powered solutions. From blockbuster deals with OpenAI and NVIDIA (NVDA) to a $130 billion backlog, Oracle's performance across the board screams momentum. This surge is fueling explosive top and bottom-line growth, while its exciting EPS outlook makes the current valuation a steal. With the stock now up almost 40% since this time last year, I remain firmly bullish on ORCL. Confident Investing Starts Here: Cloud Infrastructure Powers the AI Revolution Oracle's primary growth engine today is its Cloud Infrastructure (OCI), which is capitalizing on the AI boom sweeping the tech industry. In its most recent fiscal Q3, OCI revenue surged 51% year-over-year to $2.7 billion, outpacing major hyperscale competitors such as AWS and Azure. CEO Safra Catz pointed to 'record-level AI demand' as the key catalyst, with Oracle securing major cloud contracts with OpenAI, xAI, Meta (META), Nvidia, and AMD (AMD). These agreements contribute to a $130 billion backlog, up 62% from the previous year, underscoring Oracle's growing role in powering large-scale AI model development and deployment. Oracle's edge lies in its focus on high-performance computing and scalable infrastructure. A prime example is its deal with NVIDIA: Oracle is investing $40 billion to acquire 400,000 GB200 superchips, which will be used to build a massive AI-focused data center in Texas, expected to be operational by mid-2026. It's an ambitious move, but one that is positioning Oracle as a key player in the AI infrastructure space. Demand is strong, and with its advanced data center capabilities, Oracle appears to be just getting started. Oracle Expands Product Ecosystem with Strategy Oracle's strategy of forming strategic alliances is laying the foundation for long-term growth. A standout example is its multi-cloud partnership with AWS, which enables customers to access Oracle databases across platforms with ease. The results are already materializing—multi-cloud revenue grew tenfold year-over-year in Q3. This approach allows Oracle to collaborate with competitors while expanding its footprint in the AI space, reinforcing its leadership in hybrid and sovereign cloud environments tailored to enterprise needs. Another major initiative is Stargate, Oracle's ambitious AI infrastructure project with OpenAI and SoftBank. Oracle isn't just supplying compute power—it's helping to build a full ecosystem that supports advanced AI innovation. CEO Safra Catz also noted that bookings are converting to revenue more quickly than expected, with a 57% increase in OCI consumption revenue. These aren't just headline-grabbing deals—they're translating into real, accelerating growth. Cloud Applications Become the Unsung AI Heroes While OCI grabs the spotlight, Oracle's cloud applications, like Fusion Cloud ERP and NetSuite, are quietly powering its AI story. In Q3, cloud application revenue rose 9% to $3.6 billion, with strategic back-office SaaS applications growing 18% to $8.6 billion annually. If you are unfamiliar, these platforms utilize AI to streamline enterprise operations, from financial planning to supply chain management. For example, Fusion Cloud ERP's AI-driven analytics are deployed by companies like Walmart (WMT) to optimize workflows, aiming to boost efficiency and margins. As the chart shows, ORCL's revenues are inching higher while profit margins are sluggish, for the time being at least. What's notable about this is how these applications integrate with Oracle's broader AI ecosystem. As customers using NetSuite can tap into OCI's AI capabilities, a seamless flow of data and insights is created. This synergy drives stickiness. Once a company adopts Oracle's cloud apps, it's hooked on the ecosystem. The 16% growth in ERP revenues indicates that enterprises are betting on Oracle to modernize their operations with AI at its core. The Road to Valuation Upside Through EPS Acceleration Oracle's focus on high-margin areas like OCI and cloud applications is a recipe for margin expansion. The company's adjusted operating margin hit 44% in Q3, and with cloud revenue growing faster than lower-margin segments like software licenses (down 8%), profitability is set to climb. In the interim, consensus estimates project 8% EPS growth for fiscal 2026, accelerating to 12%, 22%, and 31% over the next three years as AI revenues scale. This trajectory shows that the market expects Oracle to convert its $130 billion backlog into high-margin cash flows. Hence, at 28x this year's expected EPS, Oracle's valuation might raise eyebrows, but I think the numbers tell a different story. The projected EPS growth, fueled by AI-driven cloud demand, suggests the stock is reasonably priced with room to run. As Oracle doubles its data center power capacity this year and triples it by fiscal 2026, revenue and earnings are set to accelerate, making today's price a bargain for long-term investors. Is ORCL Stock a Buy, Sell, or Hold? Wall Street appears relatively bullish on Oracle's prospects. ORCL stock has a Moderate Buy consensus rating, with 16 analysts currently bullish and 14 neutral. ORCL's average stock price target of $178.42 indicates a somewhat constrained upside potential of less than 1% over the coming twelve months. Oracle's AI-Powered Future Beckons Oracle's recent performance deserves serious attention from investors. This is no longer just a legacy database company—between the rapid expansion of Oracle Cloud Infrastructure (OCI), strategic partnerships, and AI-enhanced cloud applications, Oracle is positioning itself as a major contender in the AI space. With a $130 billion backlog and accelerating EPS growth projected in the coming years, the fundamentals are strong. At current levels, the stock offers a compelling blend of growth potential and attractive valuation. Bottom line: Oracle's AI journey is still in its early stages, and there could be meaningful upside ahead.
Yahoo
2 hours ago
- Yahoo
This $14 meat thermometer is perfect for the BBQ: Takes 'all the guesswork out of cooking'
Depending on its bells and whistles, food thermometers can cost over $100 — a costly investment for the casual at-home chef. Luckily, Amazon Canada shoppers don't need to spend an arm and a leg to get an "accurate" digital meat thermometer, even before Amazon Prime Day deals kick off. The bestselling ThermoPro TP03B Digital Instant Read Meat Thermometer is backed by nearly 125,000 reviews and an Amazon bestseller title. It's been dubbed a "game-changer" device by reviewers thanks to its lightning-fast high-precision sensor. More than 2,000 Amazon shoppers purchased this handy little device last month — keep scrolling to learn why. This bestselling meat thermometer has over 124,000 reviews on Amazon Canada. ThermoPro's instant-read meat thermometer detects core temperatures within +/- 0.9°F accuracy and a response time of 3-5 seconds. The 3.9" food-grade stainless steel probe is long enough to test a wide variety of meats without leaving large holes in your food. It has a backlight display for cooking in low lights or grilling at night in the dark. The thermometer auto shuts off after 10 minutes and folds away for storage and portability. It offers a temperature range of -58°F to 572°F (-50°C to 300°C), making it the perfect accessory for baking, frying, grilling and more. ⭐️ 4.4/5 stars 💬 125,000 reviews 🏆 "Game-changer." Backed by a staggering 124,000+ reviews and a 4.4-star average rating, the ThermoPro digital instant-read meat thermometer is a fan favourite among Amazon shoppers. One shopper describes the thermometer as a "great asset for your kitchen" and says that while it's "kind of cheaply constructed," it "works perfectly." This "handy" thermometer is "fantastic," writes another reviewer. It has "taken all the guesswork out of cooking." A total "game-changer." This bestselling meat thermometer can fold away easily for portability. It "does the trick," echoes a third Amazon reviewer. It's "perfect." While shoppers say the meat thermometer is "accurate" and "easy to use," some note it "does not feel sturdy" and are "concerned" about its lasting durability. Backed by over 124,000 reviews, Amazon reviewers call this "handy" digital instant-read meat thermometer an "accurate" and "easy-to-use" kitchen "game-changer." However, while it helps measure the internal temperature of meats, baking and more, some note it's "kind of cheaply constructed" and are "concerned" about its long-term durability — something to keep in mind when placing your order.


Business Insider
4 hours ago
- Business Insider
Microsoft-Backed OpenAI (MSFT) Has Reached $10B in Annual Recurring Revenue
Microsoft-backed AI company OpenAI has reached $10 billion in annual recurring revenue (ARR) less than three years after launching its popular ChatGPT chatbot. This total includes sales from its consumer products, ChatGPT business services, and its API, but does not include licensing revenue from Microsoft (MSFT) or large one-time deals, according to an OpenAI spokesperson. Interestingly, this amount jumped from about $5.5 billion in ARR last year. However, its rapid growth has come at a cost, as OpenAI lost around $5 billion last year to support its expansion. Confident Investing Starts Here: Nevertheless, OpenAI is aiming for even bigger targets. Indeed, the company is reportedly looking to reach $125 billion in revenue by 2029, according to a person familiar with the plans, which was first reported by The Information. In addition, earlier this year, OpenAI closed a $40 billion funding round that turned out to be the largest private tech deal on record and valued the company at $300 billion. As a result, based on today's revenue, OpenAI is currently valued at about 30 times sales. This highlights the huge expectations that investors have of the company. It is worth noting that OpenAI first launched ChatGPT for consumers in late 2022 and added business products the following year, with adoption growing quickly. By late March, the firm said that it had 500 million weekly active users. Moreover, earlier this month, OpenAI reported that it now has 3 million paying business users, which was an increase from 2 million in February. These strong user numbers are helping drive OpenAI's fast revenue growth as it continues to expand its AI products for both consumers and businesses. Is MSFT Stock a Buy? Although you cannot directly invest in OpenAI, you can buy shares of Microsoft, which has a 49% stake in OpenAI. And according to analysts, Microsoft stock has a Strong Buy consensus rating among 36 Wall Street analysts. That rating is based on 31 Buys and five Holds assigned in the last three months. Furthermore, the average MSFT price target of $514.93 implies 9% upside potential.