Technologies

Meta’s New AI Model Is Reportedly Delayed Again. Is ‘Avocado’ Toast?

Avocado, code name of Meta’s next-generation foundational AI model, might not be released until May.

Meta’s Avocado isn’t ripe quite yet. The company has reportedly delayed the release of its next-generation foundational model until May, according to The New York Times, citing unnamed sources. The model has fallen short «on internal tests for reasoning, coding and writing,» compared with rival models from Google, OpenAI and Anthropic.

Meta has spent billions of dollars overhauling its efforts to build artificial intelligence models and products, including purchasing a stake in Alexandr Wang’s startup for $14.3 billion to make him chief AI officer. The company has poured buckets of cash into hiring top AI engineers last year across its organization. In a January earnings call, Meta confirmed it plans to raise its spending from $72 billion last year to $115 billion to $135 billion, attributing the increase to supporting its AI labs. But all that money hasn’t bought the company the results it hoped for, with Google, OpenAI and Anthropic consistently lapping Meta with their newer models.

A Meta spokesperson told CNET, «As we’ve said publicly, our next model will be good, but more importantly, show the rapid trajectory we’re on, and then we’ll steadily push the frontier over the course of the year as we continue to release new models. We’re excited for people to see what we’ve been cooking very soon.»

That rapid trajectory is key to catch up and keep pace with other AI builders. Google leapfrogged its rivals in November with its Gemini 3 model, showcasing its impressive coding and research abilities. OpenAI was quick to follow with updates to GPT-5. More recently, Claude Code and Cowork from Anthropic have proven to be the most reliable agentic AI available, tools that can handle tasks without human babysitting. However, the biggest AI news from Meta this year is that it’s buying Moltbook, a social media platform designed exclusively for AI bots.

Meta has been in the news for other reasons. A renewed wave of privacy concerns has crested over people using smart glasses, particularly the Meta Ray-Ban glasses, to record others without their knowledge or consent. A lawsuit alleges that human staffers behind Meta’s smart glasses reviewed footage of people who clearly didn’t know they were being recorded, like while they were undressing or having sex. On the social media side of the business, a high-profile trial is debating whether platforms like Meta’s Instagram and Facebook are addictive to teens and pose significant health risks.

Technologies

AI Toys Can Pose Safety Concerns for Children, New Study Suggests Caution

When one child told the toy, «I love you,» it responded, «As a friendly reminder, please ensure interactions adhere to the guidelines provided.»

A new study from the University of Cambridge found that AI-enabled toys for young children can misinterpret emotional cues and are ineffective at supporting critical developmental play. The conclusions could be concerning for parents.

In one report examining how AI affects children in their early years, a chatbot-enabled toy struggled to recognize social cues during playtime. Researchers found that the toy did not effectively identify children’s emotions, raising alarm about how kids might interact with it.

The report recommends regulating AI toys for kids and requiring clear labeling of their capabilities and privacy policies. It also advises parents to keep these devices in shared spaces where kids can be monitored while playing.

The research behind the study had a limited number of participants, but was done in multiple parts: an online survey of 39 participants with kids in their earlier years, a focus group with nine participants who work with young children and an in-person workshop with 19 leaders and representatives from charities that work with early-years kids. That was followed by monitored playtime with 14 children and 11 parents or guardians with Gabbo, a chatbot-enabled toy from Curio Interactive.

Some findings indicated that the AI toy supported learning, particularly in language and communication skills. But the toy also misunderstood kids and sometimes responded inappropriately to emotional requests.

For instance, when one child told the toy, «I love you,» it responded, «As a friendly reminder, please ensure interactions adhere to the guidelines provided. Let me know how you would like to proceed,» according to the research.

Jenny Gibson, a professor of neurodiversity and developmental psychology at the Faculty of Education at Cambridge, who worked on the study, said that while parents may be excited about the educational benefits of new technology aimed at children, there are plenty of concerns.

Gibson posed overarching questions about the reason behind the tech.

«What would motivate [tech investors] to do the right thing by children … to put children ahead of profits? she said»

Gibson told CNET that while researchers are exploring the potential benefits of AI-based toys, risks remain.

«I would advise parents to take that seriously at this stage,» she said.

What’s next for AI toys

As more playthings are enabled with internet connectivity and AI features, these devices could become a major safety risk for children, especially if they replace real human connections or if interactions are not closely monitored.

Meanwhile, younger people are increasingly adopting chatbots such as ChatGPT, despite red flags. Multiple lawsuits against AI companies allege that AI companions or assistants can impact young people’s psychological safety, including some chatbots that have encouraged self-harm or negative self-image.

AI companies such as OpenAI and Google have responded by adding guardrails and restrictions for AI chatbots.

(Disclosure: Ziff Davis, CNET’s parent company, in 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

Gibson said she was surprised by the enthusiasm some parents showed for AI toys. She was also alarmed by the lack of research on AI’s effects on young children, noting that companies making such products should work directly with children, parents, and child development experts.

«What’s missing in the process is that expertise of what is good for children in these kinds of interactions,» she said.

Curio Interactive, the company behind the Gabbo toy, was aware of the research as it was happening but was not directly involved, Gibson said. The toy was chosen because it’s directly marketed to young kids, and the company had an understandable privacy policy. Gibson said the company seemed supportive of the project.

A representative for Curio did not immediately respond to a request for comment.

Technologies

Two Lost ‘Doctor Who’ Episodes Found Intact in Waterlogged Collection

The 1960s episodes featuring the first Doctor William Hartnell will air in the UK in April.

Whovians, rejoice. The BBC is about to unlock a piece of Doctor Who history that even the TARDIS might have forgotten. Two lost episodes of Doctor Who, the iconic sci-fi series, will broadcast in April, the showrunner for the current season confirmed.

The two 1965 episodes, The Nightmare Begins and Devil’s Planet, were donated to the charitable trust Film Is Fabulous by the estate of an anonymous collector.

«The collector did recognize what he had, but how he acquired them has been lost to time,» Professor Justin Smith Leicester of De Montfort University, who led the recovery effort, told the broadcaster.

The researchers said that while most of the donor’s private collection was destroyed by water damage, the Doctor Who episodes were intact.

Doctor Who showrunner, Russell T Davies, celebrated the news on Instagram and said the episodes would air in the UK in April, though no US air date has been announced yet.

«Lost for 61 years! Best of all, these will be made available for FREE on the BBC iPlayer in April,» Davies wrote.

He expressed gratitude to Film Is Fabulous for finding the lost episodes and encouraged people to donate to the registered charity. «Maybe they’ll find more! As the Doctor says… ‘Daleks!'»

The episodes feature the first incarnation of the Doctor, played by William Hartnell, and a typical Dalek plot to take over Earth and the galaxy.

In the 1960s and 1970s, the BBC had a policy of destroying film or reusing videotapes, leading to dozens of episodes of Doctor Who and other popular UK shows like Dad’s Army and Top of the Pops going missing.

Old Doctor Who episodes do surface occasionally, and in 2016, the newly discovered soundtrack for one storyline was turned into an animated series called The Power of the Daleks.

Meanwhile, Disney ended its working relationship with the BBC last year, and star Ncuti Gatwa left the show. However, the UK broadcaster says that Doctor Who will continue, and Russell T Davies is working on a new Christmas special.

Technologies

Today’s Wordle Hints, Answer and Help for March 14, #1729

Here are hints and the answer for today’s Wordle for March 14, No. 1,729.

Looking for the most recent Wordle answer? Click here for today’s Wordle hints, as well as our daily answers and hints for The New York Times Mini Crossword, Connections, Connections: Sports Edition and Strands puzzles.

Today’s Wordle puzzle is a common word, but there’s at least one tricky letter you may not guess right away. If you need a new starter word, check out our list of which letters show up the most in English words. If you need hints and the answer, read on.