Meta has finished pretraining its new AI model codenamed ”Avocado,” according to an internal memo.

The model is the company’s most capable pretrained base model, The Information reports, citing an internal document. Pretraining is the first phase of AI development, where a model learns basic patterns and relationships from large amounts of data.

Avocado outperforms the best freely available base models, according to the memo from Megan Fu, a product manager at Meta Superintelligence Labs. In knowledge, visual perception, and multilingual performance, it can even compete with leading fully trained models – even though Avocado hasn’t gone through the second training phase yet. This phase, called post-training, refines models for specific tasks.

Google’s parent company, Alphabet, beat Wall Street expectations on Wednesday, and is planning a sharp increase in capital spending in 2026 as it continues to invest deeply in AI infrastructure.

Alphabet on Wednesday reported profit of $34.5bn in the recently ended quarter, as revenue from cloud computing soared 48%.

The company also forecast spending between $175bn and $185bn this year, a figure much higher than analysts’ expectations of roughly $115bn.

In an earnings call, investors pressed Alphabet’s chief executive, Sundar Pichai, on the significant increase. “We’ve been supply constrained, even as we’ve been ramping up our capacity. Obviously, our CapEx spend this year is an eye towards the future,” Pichai said, in response. “We are constantly planning for the long-term.” Pichai added that he expects Google to “go through the year in a supply-constrained way”.

Apple Intelligence researchers are proposing a new approach to text-to-speech that would make Siri quicker to respond. That might, though, also make conversations flow more naturally.

Apple may have lost the odd AI researcher, but it continues to publish significant papers on the topic. Previously it has published about limiting AI from taking actions a user didn’t approve, and examining how to prevent hallucinations.

Now in a study called ”Principled Coarse-Grained Acceptance for Speculative Decoding in Speech,” researchers from Apple and Tel Aviv University, have focused on text-to-speech applications.

In AI, speech is sometimes generated based on tokens, or very short samples of sound. These are phonetic sounds, measured in milliseconds, which are then assembled into sentences.

A leaker with only a mixed track record is doubling down on the rumor that the iPhone 18 Pro is not getting a major redesign, but now says it’s because the iPhone 17 Pro was so successful.

Back in December 2025, leaker ”Fixed Focus Digital” said that the iPhone 18 Pro wouldn’t see visible changes ”as large as everyone thought.” Now he or she is back on Chinese social media site Weibo to say exactly the same thing, but with one qualification.

In translation, the leaker describes the iPhone 17 has having been enough of a hit for Apple to be celebrating with champagne. Then Fixed Focus Detail that the iPhone 18 series will not change much.

He or she does then go on to say that the A20 and A20 Pro processors expected to be in the iPhone 18 range will be significant. Or at least that they will be the most significant part of the update.

YouTube reported more than $60 billion in revenue for 2025.

The new disclosure covers revenue across advertising as well as subscriptions for services such as YouTube Premium TV, and gives somewhat of a window into how massive the userbase for the service is. In total, the company reported 325 million paid subscriptions in its fourth quarter across consumer services, which includes YouTube Premium, as well as the paid cloud storage service Google One, up from 300 million in the prior quarater.