The rivalry between OpenAI and Google has intensified since the launch of ChatGPT, with tensions particularly evident leading up to this week’s I/O conference. Google employees expressed concern that Sam Altman might attempt to overshadow their event as he did last year when OpenAI unveiled an advanced voice mode for ChatGPT the day before Google’s showcase.
This time, OpenAI announced significant news the day after the conference.
OpenAI’s acquisition of the “io” hardware division from Jony Ive’s design firm, LoveFrom, has been interpreted by some as a calculated move to capture attention. While the division’s name supposedly stands for “input output,” the strategic timing of the announcement diverted focus from Google’s strong presence at the I/O event. This exchange illustrates the dynamics between the two leading AI companies: while Google boasts technically advanced and widely used models, OpenAI excels in brand presence and public engagement.
At the heart of OpenAI’s recent announcement is a $6.5 billion equity deal aimed at hiring around 55 individuals from LoveFrom, including former Apple design leaders Evans Hankey, Tang Tan, and Scott Cannon. These new hires are set to report to Peter Welinder, an established product leader at OpenAI, directly under Altman’s supervision. Other LoveFrom designers, including notable figures like Mike Matas, will remain with Ive, who is currently working on the first electric Ferrari and advising Airbnb CEO Brian Chesky. According to OpenAI’s press release, Ive and LoveFrom will take on significant design and creative roles within the organization.
Once LoveFrom completes its existing contracts, Ive and his team will concentrate exclusively on OpenAI while maintaining their independence. Meanwhile, OpenAI is actively seeking to fill roles related to the future of computing within the team it has acquired. One job listing invites candidates who have experience “teaching models to speak and perceive,” offering a compensation package ranging from $460K to $555K along with equity.
The timeline leading to this acquisition is noteworthy: Altman and Ive initially met two years ago and formally decided to collaborate on hardware approximately a year ago. The io division was established at LoveFrom with a small team of OpenAI employees. Towards the end of 2024, OpenAI and Laurene Powell Jobs invested in this initiative, amid discussions to raise substantial funding for its establishment as an independent entity. It is also notable that OpenAI’s startup fund, which is not owned by the main company, participated in this financial backing.
Ive ended a lucrative consulting relationship with Apple in 2022, which had restricted his ability to compete with Apple’s products. Now, he and Altman are positioning themselves to potentially unveil a voice-first AI device in the coming year.
Early models of this upcoming device are already in existence. Altman recently noted in a promotional video with Ive that the device will be portable, capable of being placed on a desk or worn around the neck. Supply chain sources have indicated that it could be comparable in size to an iPod Shuffle.
In addition to the forthcoming device, Ive and Altman are also exploring the development of AI earbuds. Altman aims to integrate hardware options with ChatGPT subscriptions, envisioning a range of AI-driven products designed to reduce the company’s dependency on Apple and Google for distribution. With his deal with Apple behind him, Ive seems poised to challenge the company he once helped create.
In contrast, Google showcased its advancements this week with the widespread rollout of AI Mode in Google Search. Although the company’s product strategy appears less unified compared to OpenAI’s, it is beginning to utilize its extensive personal data reserves to enhance the functionality of its Gemini AI. By integrating Gemini with platforms like Gmail, Workspace, and YouTube, Google aims to retain its user base in a manner reminiscent of Meta’s integration of Stories in Instagram, which effectively deterred users from migrating to Snapchat.
Following interactions with various Google employees during I/O, it became clear that the company does not foresee an imminent crisis, despite external speculation. While there is an awareness that buying out distribution channels on Apple devices may soon be more challenging, Gemini is approaching 500 million monthly users.
While ChatGPT is certainly encroaching on Google Search’s territory, the company has demonstrated a surprising agility in updating its search capabilities. This contrasts sharply with Apple, which is lagging in AI competitiveness and experiencing internal disputes that Google has largely resolved over recent years.
There is no doubt about Google’s strong position in AI model development. The latest iterations of Gemini are impressive, and the company is strategically aligning its AI initiatives to adapt to a post-phone landscape through its Project Astra. Moreover, Google has the computational power to roll out advanced tools like the new Veo video model, unlike OpenAI’s Sora, which remains severely limited by GPU access.
Even with ChatGPT’s outstanding growth, there remains uncertainty about whether Gemini will gain significant household recognition. If this dynamic persists, it could signal a fundamental shift in technology usage that might negatively impact Google’s future. For the moment, however, it appears that Google is poised to maintain its footing.
Elsewhere
- Anthropic made headlines this week as it unveiled its Claude 4 models, which it claims are the best available for coding. With key players like OpenAI, Google, and Meta competing for dominance in AI interfaces, Anthropic is positioning itself as a model provider of choice. Notably, Windsurf, a company in discussions to sell to OpenAI, appeared to be deliberately excluded from immediate access to the new models, leading to speculation about industry dynamics. “If models are countries, this is the equivalent of a trade ban,” Nathan Benaich remarked on X. (Additionally, one must question the implications of releasing models known for potentially dangerous applications.)
- Microsoft Build faced interruptions from protests as it sought to showcase exciting news, including the launch of Elon Musk’s Grok model on Azure and innovations in AI plumbing. Unfortunately, the protests focusing on the company’s dealings with Israel overshadowed these developments, leading Microsoft to attempt (unsuccessfully) to restrict the use of certain terms related to the conflict in internal communications.
- I recently tested a prototype of Google’s smart glasses. During a brief trial, I experienced a low-resolution display providing voice interactions with Gemini, basic Google Maps functionality, and photo capture. The technology remains in its early stages, and Google acknowledges that full AR capabilities are still years away. In partnership with brands like Warby Parker and Kering, the company is preparing to release Android XR-enabled glasses next year, and with Apple planning similar offerings in the near future, Meta’s early lead in this space may be narrowing.
Personnel log
- YouTube has appointed Justin Connolly from Disney to lead its media and sports team, a transition that Disney is currently contesting legally.
- Tinder’s CEO Faye Iosotaluno is stepping down, with her responsibilities shifting to Match Group CEO Spencer Rascoff.
- Vladimir Fedorov, formerly with Meta’s engineering team, has joined GitHub as CTO.
- Will Robinson, previously the VP of engineering at Coinbase, has joined Plaid as CTO.
- Stephen Deadman, Meta’s VP of data protection in Europe, is departing amidst “structural changes.”
Link list
More to explore:
- My exclusive hands-on experience with Google Beam — a groundbreaking demo.
- Demis Hassabis and Sergey Brin discuss AGI at I/O.
- Elon Musk’s tense appearance with Bloomberg.
- A look at the demand for AI research talent.
- Tech CEOs are utilizing AI to handle earnings calls.
- Insights into ByteDance’s financial strategy.
- Telegram reported profitability last year and targets $2 billion in revenue this year.
- “Crypto elite rattled by ‘wrench attacks’ and other criminal events.”
- A recommended read: “Empire of AI” by Karen Hao.
If you haven’t already, consider subscribing to Technology News for unrestricted access to Command Line and all our reporting.
As always, feedback is welcome, whether you have insights on this issue, opinions on simulations, or story ideas to share. You can respond here or contact me securely via Signal.
Thank you for your continued support.
