Squeezed into the remodelled Google Berlin old medical school theatre I have a lovely glass of wine ready to go as I attempt to live blog a keynote for the first time in a while. I love watching American keynotes with Europeans, as we are much more… sceptical. One person in the crowd cheered for an announcement once and everyone looked at them like they were crazy.
OK, this is all pretty unstructured as I don’t have time to construct a full blog post right now. here we go!
AI, AI, AI, and yes… AI
Google announced that they are doubling down on AI and it’s “making the world better”, but is it? We’re on the cusp of a new era that helps everyone do “better”, but it’s hard to feel like that right now when AI is causing layoffs, downturns, removing the opportunity for people’s creativity etc.
“Seven years into our journey as an AI first company”.
Where did that come from? Google are re-imaging all projects and products, will all the AIs that start communicating with each other to make our lives “easier” start ignoring each other?
All key Google apps now getting AI makeovers. Gmail gets some crazy improvements to autocompletes that feel a bit impersonal to me really. Maybe there will be a new movement of pride for people who don’t use AI assistants. Maps… The maps generated immersive view does look good, but rather overly perfect. I think people might end up disappointed with the reality when they hit the ground.
Google photos will let you edit photos, create your perfect pictures, make everything perfect, hmm… Will this be “deepfake as a service”?
“Making AI helpful for everyone”
Is the new slogan it seems. What will happen when AI jumps the shark? Will it ever? Then what will Google do?
New PaLM 2 LLM, it has a Gecko version for mobile and offline use, and different models for different uses, such as medical. Also used in Bard.
Deepmind makes a comeback, but what about tensorflow? Very little mention of it.
Water marking in models and generated content and source material with embedded metadata. How will they ensure it won’t be removed? Someone told me on Mastodon how that might work
Bard generation can now export from generated sources and models into other tools. Plus using sources from other Google apps and sources of knowledge inside of Bard itself. This includes search, so really Bard will be the tool that replaces everything? Combining Low code, AI, office tools, exporting etc… It’s impressive, but is it all too much? Partnerships with other competitors are interesting. But I guess this is what Google marketplaces have always done.
I have a note about project Gemini, but can’t remember the details now, so here’s an article about it from elsewhere.
Google Workspace files will now have prompts for creating templates, words are becoming more and more important for generating content. It’s called Duet AI for workspace, so now we’re paying for AI finally.
For example, creating stories together in a doc with an assistant, “Sidekick”. It will “cite” sources it uses for results, does it keep up to date as things change? Though some of the examples felt forced, or “toy”, but we’ll see.
Then we get to how AI changes Google search. Hammering home the source of information, will those sources matter? What do the sources think about being cited? Taking more away from them? It took 45 minutes for Google to mention Ads, will they continue to be a core part of their business? How will they pay for everything without them?
I don’t see search looking that much different from how it already does really, will people use it any differently? It’s more like faceted search really… Just fancier. I will continue to use Duck Duck Go in Safari personally 😆.
Actual developer content
Google cloud… Vertex AI will present new models people can use in their own products. More partnerships again, “imagen”, “codey” based on your own codebase, Chirp.
We’re helping you “seize this moment”
But that’s left for people to do?
A3 virtual machines based on Nvidia. If other people are monetising AI, Google wants chasing the hosting of it.
PaLM has an API, go wild developers.
Project Tailwind for note taking… How does Tailwind CSS feel?
Responsible and bold
Pushing the concept of responsible a lot, but if they are taking jobs and creativity away from so many, how is that responsible? Didn’t they fire a lot of the ethics team?
“Guardrails”, only with authorised partners. And who are they? How do we know they are adhering to them?
Reducing problematic outputs by using automated adversarial testing.
Android
Almost as an after thought, but still plenty of AI of course.
Android now supports Multitasking, tablets, and foldables better
Android works with the “find my network”, awkward swallow… Shows Apple devices too.
Another RCS push… No one cares outside the US, despite claiming a potential 800 million users. Maybe that’s “uses”?
Emoji wallpaper, ho hum.
At this point I left as my head was spinning, so I report now from third party sites.
And then a foldable. I actually like the form factor, but way too expensive for me.
Pixel tablet. I tried Anroid tablets in the past, but have become an ipad person in recent years. The main issue has always been apps that take advantage of Android tablets. We will see.
Google AI
And that’s about it, or where I lost track anyway. Some impressive technology and announcements, but is this also a watershed moment were AI really truly starts to take over all our work and creative pursuits? Wether we like it or not? That is a scary prospect.