22 Comments
User's avatar
Pud's avatar

I’m now thinking the io is going to be like the very first iPod Shuffle (the white one that Steve Jobs likened to a thin pack of gum) that even had a lanyard you could wear around to your neck! Jony Ive going for that early 2005 nostalgia!

Expand full comment
Francis Harvey's avatar

Your crazy thoughts inspire me to imagine that the 'io' device could turn out to be super Ive-designed Alexa update. Just a thought... But maybe it will be an 'earthquake', maybe it will open a portal to a new universe. Curious I am, but I can wait again for another earth-changing IT development. I survived the Windows 2.0 introduction after all. I'm ready!

Expand full comment
Alberto Romero's avatar

The no-screens part makes me wonder if this could be a good thing after all, compared to what we have.

Expand full comment
Bruce Spector's avatar

here’s your invite to the Vibeathon https://partiful.com/e/dFWMaelGpf0HW7SUwq6J

Expand full comment
Bruce Spector's avatar

hundreds of rsvp’s to the attap.ai Vibeathon event at upcoming #NYTechWeek strongly suggest people are craving the return of agency and control to the individual. see https://attap.ai/vision/

Expand full comment
Bruce Spector's avatar

i’d prefer no control for THEM and complete control for US.

Expand full comment
Nicolai Jepsen's avatar

For some reason I keep thinking of IO as some sort of Pokédex. A device you could point at something, that could explain what you’re looking at, experiencing, and so on. Would love to reduce core devices, though. One “life device,” (currently iPhone) and one “work device,” Mac.

Expand full comment
Thrawn's avatar

If there's a business to be made in ditching a screen completely, Google will be able to do fine there. Yes it took Google a year or two to make Android multi touch as good as iPhone but it quickly surpassed it and has never looked back. Could they beat Google to a screenless AI device? No, Google already has many of those. Could they make a more ergonomic one? Undoutably, but Google won't be far behind

Expand full comment
Alberto Romero's avatar

I don't think Google can afford to risk its empire on that bet

Expand full comment
Thrawn's avatar

It doesn't need to, it already has Gemini on Bluetooth speakers with a camera. Putting a battery and a Jony Ive stamp on it isn't a dramatic change. Whatever money there is to make, they'll be ready

Expand full comment
Alberto Romero's avatar

That's a feature. We're talking here about changing the entire substrate on which modern computing and interaction happens. Just like the iPhone did in 2007. Where would they get their money?

Expand full comment
Thrawn's avatar

Concretely we're talking about a Pixel Phone without a screen. Gemini is the leading mutimodel model, and it already can do things like call a store to check if they're open for you without you needing to do anything. They showed it buying something for you without needing to open a website. They'll make the money the same way OpenAI will. Yes that will be at the expense of their advertising business, but that is true regardless of if they react or not. And they're already reacting. Gemini is where Siri and OpenAI would love to be, and I don't see them catching up.

Expand full comment
Thrawn's avatar

I start every day asking Gemini to summarize the last 24 hours of my emails and documents and tasks and it doesn't care that I am not looking at all the advertisement in my email. They know the change is coming and they're reacting super quick imo

Expand full comment
Alberto Romero's avatar

I don't think Google makes much money from email add tbh

Expand full comment
Joseph de Castelnau's avatar

While Google I/O had a lot to offerand your article is right on for Altman's ambition (assuming Ive figure the next deviceless iphone) I caution against Sundar's announcement. While Google is back in play, I see a pattern emerging of all these companies solutioning with AI to roach motel their leads. This is a dangerous pattern for humanity. Do we really need an AI assistant to help us craft anticipatory email instead of having proper data retrieved while we email. That one pushes too far in order to keep their moats. Others are better. When do we talk about these type of metrics. Replit is the same way: Dangerous (unless used as just a prototype but then again you can do that in gemini). Where are those metrics. Honestly the llarena is pretty lame as as success metrics. We as an industry can do so much better than that. When the reward functions are misaligned AI hurts us. When they are not, we are great digital assistants, decision support tools. Please mention this in your usually fair and good coverage.

Expand full comment
Mike's avatar

Nice write up! You hit in a key belief. But you might consider they aren't just building AI, they are trying to build utopia. Look closely at what the talk about and you find what they really care about.

Beyond that, anything OpenAI can create, can be copied within a short time.

Google has the many hands.

It's always about funneling the "much" into something useable... Which they make a lot of mistakes on. It's the problem being spoiled with a wealth of very good ideas, which one will be great.

Expand full comment
Jatinder Bali's avatar

One of the best articles online. I am sure it took you time to pen it down and kudos to you.

Expand full comment
Alberto Romero's avatar

Thank you Jatinder!! It took me two full days actually, so around 16 hours. Much more than usual that's for sure

Expand full comment
Shaeda Ramblings's avatar

Do we have any kind of metrics on how Gemini's 2.5 models impacted OpenAIs?

Or is ChatGPT somehow still growing despite being the inferior product?

Expand full comment
Alberto Romero's avatar

Hmm ChatGPT is not "inferior" strictly speaking. They're more or less comparable. But Google's ecosystem is much broader and deeper. Yo answer your question, ChatGPT is still growing very fast!

Expand full comment
User's avatar
Comment deleted
2dEdited
Comment deleted
Expand full comment
Alberto Romero's avatar

Haha I never use ChatGPT for a first draft holy shit that would make them terrible! The earthquake idea is mine and also the "you made trash" bit (ChatGPT is too polite to suggest that). This one I actually wrote in almost one pass in English (except the Google part that I had to get all the references and a couple editing passes). So no Spanish and no ChatGPT involved (except probably some isolated words here and there)

Expand full comment