Forget the old Apple slogan, "Think different." For Deepak Singh, VP of developer agents and experiences at AWS, the mantra of the future is "work differently," and the way he wants to do that is through agentic AI. "I think people get too hung up on the automation and efficiency, part of which are outcomes," said Singh. "We are working differently, but the way we are working different is making us more effective because [agents are] solving harder problems or more problems than you could do before." Singh sat down with ZDNET on Wednesday, shortly after AWS introduced a bevy of new tools and features centered around agentic AI solutions. Among the biggest announcements were Amazon Bedrock AgentCore, a new enterprise-grade platform designed to facilitate the implementation process for new agents, and a new virtual store within AWS Marketplace, which allows customers to choose agents from Anthropic, IBM, Perplexity, Salesforce, and other vendors.

Looking for an automated way to record, analyze, and summarize meetings and other conversations? You can now do that with any paid ChatGPT subscription, as long as you use a Mac. Also: Is ChatGPT Plus really worth 20 when the free version offers so many premium features? On Wednesday, OpenAI announced that its new Record mode is now accessible to ChatGPT Plus subscribers. The recording capability, which kicked off in June, was initially restricted to the more expensive ChatGPT Pro, Enterprise, and Edu plans.

Over the last year, chain of thought (CoT) -- an AI model's ability to articulate its approach to a query in natural language -- has become an impressive development in generative AI, especially in agentic systems. Now, several researchers agree it may also be critical to AI safety efforts. On Tuesday, researchers from competing companies including OpenAI, Anthropic, Meta, and Google DeepMind, as well as institutions like the Center for AI Safety, Apollo Research, and the UK AI Security Institute, came together in a new position paper titled "Chain of Thought Monitorability: A New and Fragile Opportunity for AI." The paper details how observing CoT could reveal key insights about a model's ability to misbehave -- and warns that training models to become more advanced could cut off those insights. A model uses chain of thought to explain the steps it's taking to tackle a problem, sometimes speaking its internal monologue as if no one is listening. This gives researchers a peek into its decision-making (and sometimes even its morals).

Just a year and a half ago, the latest and greatest of Adobe's Firefly generative AI offerings involved producing high-quality images from text with customization options, such as reference images. Since then, Adobe has pivoted into text-to-video generation and is now adding a slew of features to make it even more competitive. Also: Forget Sora: Adobe launches'commercially safe' AI video generator. On Thursday, Adobe released a series of upgrades to its video capabilities that give users more control over the final generation, more options to create the video, and even more modalities to create. Even though creating realistic AI-generated videos is an impressive feat that shows how far AI generation has come, one crucial aspect of video generation has been missing: sound.

Hackers are stashing malware in a place that's largely out of the reach of most defenses--inside domain name system (DNS) records that map domain names to their corresponding numerical IP addresses. The practice allows malicious scripts and early-stage malware to fetch binary files without having to download them from suspicious sites or attach them to emails, where they frequently get quarantined by antivirus software. That's because traffic for DNS lookups often goes largely unmonitored by many security tools. Whereas web and email traffic is often closely scrutinized, DNS traffic largely represents a blind spot for such defenses. Researchers from DomainTools on Tuesday said they recently spotted the trick being used to host a malicious binary for Joke Screenmate, a strain of nuisance malware that interferes with normal and safe functions of a computer.

SAVE 250: As of July 17, the Eufy C10 robot vacuum is on sale for 229.99 at Amazon. This is 52% off its list price of 479.99 and marks its lowest-ever price. If you're feeling like you may have missed your chance to score a robot vacuum on sale during Prime Day, Amazon thankfully still has a few models available sporting some excellent discounts. One of our favorites that has caught our eye is the Eufy C10 robot vacuum, which is still at its lowest-ever price. The Eufy C10 robot vacuum is usually listed for 479.99, but it's currently marked down to 229.99.

Roblox is a popular digital space for kids and teens to congregate while playing their favorite video game, or in the platform's parlance, experience. Now teens ages 13 to 17 who want to access a special feature designed to make those hangouts even more fun will have to verify their age via a video selfie. Roblox announced the new requirement Thursday as part of a slate of safety and privacy measures. Once Roblox estimates the user's age -- via the AI-powered age verification product Persona -- and assigns a qualifying age group to their account, it allows them to take advantage of the new feature, called "Trusted Connections." Teen users can add each other as Trusted Connections, which allows them to communicate via voice and chat without filters.

A stunning Miss England semi–finalist has been revealed, but there's a huge catch – she is AI–generated. The Miss England pageant has launched a brand new AI round, featuring computer–generated beauty queens. Glamorous contestants can now walk down a virtual catwalk by making digital twin avatars of themselves as part of the Black Mirror–style qualifying round. Organisers believe they are the first beauty pageant in the world to introduce a digital AI round which will help'reflect the world' the young women are stepping into. The contest has already axed its bikini swimwear round in a bid to move away from outdated pageant stereotypes and replaced it was a CPR round teaching life–saving skills.

Earlier this week, Grok unveiled two animated AI companions for users to interact with, one of which is a flirty anime girl who puts on lingerie if you compliment her enough. Now the AI chatbot is adding a third character to the mix, only this time it's apparently modelled on what xAI CEO Elon Musk thinks women find attractive. Musk introduced Grok's new male AI companion in an X post on Tuesday, asking for suggestions in naming the character. Illustrating the request was a picture of the new companion: a stern-looking anime man wearing a black suit. "His personality is inspired by Edward Cullen from Twilight and Christian Grey from 50 Shades [of Grey]," wrote Musk.

What makes an AI browser better than a regular browser? With Perplexity's Comet, that's what I'm trying to find out. Last week, AI search engine company Perplexity launched its highly anticipated browser, which uses the Perplexity search engine and comes with a built-in AI assistant. As AI models become more advanced with chain-of-thought reasoning for handling complex tasks, they're capable of more agentic, or autonomous, work. This is a natural fit for web browsers, which are essentially the gateway to our online lives.