However a basis mannequin’s popularity today rests largely on its coding capabilities, and for months Google’s coding instruments have been outgunned by Anthropic’s Claude Code and OpenAI’s Codex. These methods are so dramatically superior to Google’s personal choices that the corporate has reportedly needed to permit some engineers at DeepMind, its AI division, to make use of Claude for his or her work—lest they fall farther behind.
So once I arrive on the convention in Mountain View, California tomorrow, I’ll actually be looking out for any efforts Google is making to claw its approach again into frontrunner place. However I’m additionally desperate to see new developments in areas the place Google shapes the leading edge, corresponding to AI for science. The corporate’s strikes there may obtain much less consideration, however they are going to be no much less consequential.
Listed below are three issues I’ll be paying explicit consideration to over the following two days.
An tried coding comeback
Google is taking its AI coding disaster severely. In accordance with reporting from The Information, there’s a brand new AI coding staff at DeepMind. And the Los Angeles Times has reported that John Jumper, who shared a 2024 Nobel Prize in chemistry with DeepMind CEO Demis Hassabis for his or her work on the protein construction prediction software program AlphaFold, is lending his abilities to the efforts. I might be stunned if we don’t see a serious new coding launch at I/O, maybe within the type of an replace to the corporate’s Antigravity agentic coding platform.
That stated, we shouldn’t count on something transformative right here. Googlers have entry to fashions and merchandise which can be considerably forward of these launched to the general public, but they had been nonetheless reportedly combating over who obtained entry to Claude Code final month. Except the corporate has made astonishing progress since then, Google in all probability received’t make it again to the coding frontier within the subsequent two days.
Science and well being
Coding could be Google DeepMind’s weak spot, however science is its conspicuous energy. It’s the solely frontier AI firm to have earned a Nobel Prize. And as LLMs have come to dominate the AI-for-science landscape, Google has solely solidified its lead. Final 12 months, the corporate launched a number of scientific AI instruments, together with the AI co-scientist, which formulates hypotheses and analysis plans in response to person questions and has been described as an “oracle” by one Stanford scientist, and AlphaEvolve, a system that iteratively discovers new options for mathematical and computational issues. If any new scientific instruments are introduced at I/O, they’ll be price noting.
I’ll even be paying shut consideration to any strikes Google makes in well being and drugs. Google is doing a number of the greatest analysis on the market on LLM-based health tools, however OpenAI has outlined the well being AI dialog because the launch of ChatGPT Well being in January. Google has announced that it will likely be making its AI-powered Well being Coach publicly obtainable tomorrow, however promotional materials means that the instrument is geared extra towards offering recommendation on subjects corresponding to health and food regimen than to addressing customers’ medical issues. Is that this one other space the place Google has fallen behind, or is the corporate exercising acceptable warning in a high-stakes area?
The drama
Whereas Google followers congregate down in Mountain View, roughly 30 miles north in Oakland the Elon Musk v. Sam Altman trial will probably be wrapping up. The previous few months have seen greater than their justifiable share of AI CEO drama—earlier than the trial, the animosity between Altman and Anthropic CEO Dario Amodei took middle stage as Anthropic and OpenAI labored to barter offers with the US Division of Protection. However DeepMind’s Hassabis has, for essentially the most half, steered away from such drama. He successfully presents himself as a Nobel Prize-winning nerd, and if he has written screeds about any of his friends, they haven’t been leaked to the press or appeared in authorized discovery.
That’s to not say that Google is controversy free. Final month, a bunch of 600 staff, lots of whom work for DeepMind, despatched a letter to CEO Sundar Pichai protesting an impending DoD deal. Google signed that deal the following day. Hassabis, Pichai, and all the opposite massive names will certainly do their greatest to skirt these and different sensitive topics whereas on stage, however controversies will worm their approach in regardless. It will likely be attention-grabbing to see whether or not Google can preserve its veneer of neutrality.
