4 ms·

Around 2013 is when I/Os became less about showing technological progress and more about corporate talk. Examples of I/O topics before 2013: Android, Maps API,

by clnq 3y ago

Around 2013 is when I/Os became less about showing technological progress and more about corporate talk. Examples of I/O topics before 2013: Android, Maps API, App Engine (cloud), Chrome, AJAX for their APIs, real-time collaboration/Wave, Google TV (early predecessor to Android TV), Chromebooks, Chrome Web Store, Google Drive, Google Docs, Google Glass, Gmail updates, Google+, YouTube updates. Most I/O topics after 2013 revolved around marginal improvements on a lot of things, like individual Android features and apps. And naturally, when there is so little to talk about, corporate speak fills the void. There were still some interesting topics after 2013, especially if you follow some specific developments in the Android ecosystem. But there were very few groundbreaking topics. Google will probably talk about Bard this year, which has not been a huge success, but at least it's something bigger that harkens back to pre-2013 days.

kordlessagain 3y ago
Bard needs API access.
- verdverm 3y ago
  There is Vertex AI, which some people are getting early access to (what i hope is) PaLM2 | ULM. The GCP offering only contains tools for your own models, what everyone wants is an alternative to ChatGPT, a pre-trained LLM from Google accessible via API and fine tune-able in Vertex. There is this in the meantime: https://github.com/acheong08/Bard https://github.com/acheong08/Bard
  - verdverm 3y ago
    This aged well :] APIs are available since the middle of the presentation!
- dragonwriter 3y ago
  They are providing API access to at least some models in the same family underlying Bard, but explicitly (while its in preview) prohibiting production/commercial use — not disclaiming liability and advising against it as is common for pre-release products, but actually prohibiting in the ToS. Google is taking an hypercautious approach around “AI safety” issues that amounts to “we should not release products and if we do they are just demos that no one should be permitted to use for any serious purpose” which is going to hurt them in the marketplace but also does nothing meaningful to deal with either the real and immediate issues with AI or the mixture of other (some real but less immediate but largely science fantasy) issues that typically are referred to under the “AI safety” umbrella (and that’s even more true if you consider from the pov of assuming that the more science-fantasy-ish of those risks are real and serious.)
vesinisa 3y ago
Almost all of the topics you listed have sessions in this year's Google I/O. There are 35 different sessions tagged "Android" alone: https://io.google/2023/program/?q=android https://io.google/2023/program/?q=android There might be no sessions for stuff like how to use web APIs because these are already mature and well-documented technologies. I/O is about showcasing new stuff after all.
- clnq 3y ago
  I don't know, it seems to me that recently in I/O, discussions have primarily focused on mature core products, indicating a shift towards iteration rather than invention. I would say the conference definitely showcases less "new stuff" or "new stuff" of smaller scope. Both I/O and Google now seem to prioritize safe, low-risk technologies from a business standpoint. And that just doesn't fill up an entire conference without some business speak padding. To my mind, the turning point towards this new, less fun and more corporate Google and I/O was around 2013. I think that's when they started thinking about shaping the company into Alphabet with Google as a subsidiary as well. Famously they removed "don't be evil" as the motto and expanded into hardware a lot soon after 2013, too. It was a time of wholistic change at Google.
gjvc 3y ago
RIP Google Wave, we hardly knew ya.
- clnq 3y ago
  It was kept alive internally by Apache for a while later — https://incubator.apache.org/projects/wave.html https://incubator.apache.org/projects/wave.html. Very interesting why it failed as a project in two separate companies. There was an explosion of team collaboration software in the 2010s and yet these companies couldn't make their product compete.
  - gjvc 3y ago
    "The Wave project retired on 2018-01-15."
- shmoe 3y ago
  Wave was awfully confusing. What was it even for?
  - clnq 3y ago
    It was a more limited Notion.
derefr 3y ago
> Most I/O topics after 2013 revolved around marginal improvements on a lot of things, like individual Android features and apps. And naturally, when there is so little to talk about, corporate speak fills the void. I don't understand how a corporate culture that's so focused on shipping new projects at the expense of maintenance, manages to ship so few new projects.

by clnq 3y ago

kordlessagain 3y ago

Bard needs API access.

verdverm 3y ago
There is Vertex AI, which some people are getting early access to (what i hope is) PaLM2 | ULM. The GCP offering only contains tools for your own models, what everyone wants is an alternative to ChatGPT, a pre-trained LLM from Google accessible via API and fine tune-able in Vertex. There is this in the meantime: https://github.com/acheong08/Bard https://github.com/acheong08/Bard
- verdverm 3y ago
  This aged well :] APIs are available since the middle of the presentation!
dragonwriter 3y ago
They are providing API access to at least some models in the same family underlying Bard, but explicitly (while its in preview) prohibiting production/commercial use — not disclaiming liability and advising against it as is common for pre-release products, but actually prohibiting in the ToS. Google is taking an hypercautious approach around “AI safety” issues that amounts to “we should not release products and if we do they are just demos that no one should be permitted to use for any serious purpose” which is going to hurt them in the marketplace but also does nothing meaningful to deal with either the real and immediate issues with AI or the mixture of other (some real but less immediate but largely science fantasy) issues that typically are referred to under the “AI safety” umbrella (and that’s even more true if you consider from the pov of assuming that the more science-fantasy-ish of those risks are real and serious.)

vesinisa 3y ago

Almost all of the topics you listed have sessions in this year's Google I/O. There are 35 different sessions tagged "Android" alone: https://io.google/2023/program/?q=android https://io.google/2023/program/?q=android There might be no sessions for stuff like how to use web APIs because these are already mature and well-documented technologies. I/O is about showcasing new stuff after all.

clnq 3y ago
I don't know, it seems to me that recently in I/O, discussions have primarily focused on mature core products, indicating a shift towards iteration rather than invention. I would say the conference definitely showcases less "new stuff" or "new stuff" of smaller scope. Both I/O and Google now seem to prioritize safe, low-risk technologies from a business standpoint. And that just doesn't fill up an entire conference without some business speak padding. To my mind, the turning point towards this new, less fun and more corporate Google and I/O was around 2013. I think that's when they started thinking about shaping the company into Alphabet with Google as a subsidiary as well. Famously they removed "don't be evil" as the motto and expanded into hardware a lot soon after 2013, too. It was a time of wholistic change at Google.

gjvc 3y ago

RIP Google Wave, we hardly knew ya.

clnq 3y ago
It was kept alive internally by Apache for a while later — https://incubator.apache.org/projects/wave.html https://incubator.apache.org/projects/wave.html. Very interesting why it failed as a project in two separate companies. There was an explosion of team collaboration software in the 2010s and yet these companies couldn't make their product compete.
- gjvc 3y ago
  "The Wave project retired on 2018-01-15."
shmoe 3y ago
Wave was awfully confusing. What was it even for?
- clnq 3y ago
  It was a more limited Notion.

derefr 3y ago

> Most I/O topics after 2013 revolved around marginal improvements on a lot of things, like individual Android features and apps. And naturally, when there is so little to talk about, corporate speak fills the void. I don't understand how a corporate culture that's so focused on shipping new projects at the expense of maintenance, manages to ship so few new projects.