The problem is that when you and three kooks say that Google is bad, people are generally not able to tell the difference between the kooks and you, even if what you are saying is quite sensible. Sigh.
Or it could go the way of people's dreams for mainstream consumer crypto in 2016.
Every business desperate for a boost in cash (and a pool of desperate LinkedIn posters) are pretending that they just invented machine learning in the last 18 months.
"AI" has become synonymous with either "snake oil" or "desperate marketing".
That might help pump the stocks of these companies, but government picking winners isn't necessarily a good thing for consumers.
The government has sunk billions into EV incentives for instance, but consumers are currently struggling with inflation for essentials like food and shelter. A slightly cheaper EV isn't really a priority for many, which explains why EV inventories are piling up.
Immediate, short-term perception of value doesn’t matter for long-term strategy.
I suppose we’ll see in 2-3 years how many businesses maintain their newly found focus on “AI”. I’m betting that the organizations that aren’t employing internal teams training their own models will drop off quickly.
A lot of companies are just throwing together an interface to OpenAI and praying for the best. Long-term this won’t work. Users will eventually demand 100.0% data accuracy.
The current valuation of Bitcoin has nothing to do with whether business models attempting an inclusion or pivot to crypto was successful because most businesses completely dropped their attempts.
The downsides to having a golden goose are that a silver goose just can't compete, another golden goose is a threat. Golden geese are rare too and so much of it is luck. For example, Google's search and ads complete dominance began in an era when nobody cares about search. They considered the problem "solved" and AltaVista was peak technology.
The paper that set off the LLM gold rush was written in 2017 by 8 Google researchers in 2017, none of whom still work at the company. Can you imagine where Google might be if they embraced this 7 years ago?
Instead, they're making "me too" announcements taht so reminds me of Microsoft with Bing or Meta with, well, just about anything these days (particularly IG and Tiktok "me too" features).
Google has other golden geese but almost all of them are acquisitions (eg Youtube, Maps, Android, Docs, Youtube). Chrome is the notable stand out.
Going through this list the first 53 or so are AI "me too" announcements. Google has blown a huge potential AI lead and has become an un-agile reactive behemoth, possibly even a dinosaur.
Notable that Google I/O is a developer conference. Yet it feels like this is more about "look how we crammed some AI into our products" than "here's all the amazing new capabilities you developers can start tapping."
The thought of Google projects only fills me with dread. They move into an area. Crush the competition with the market advantage, then abandon their products, leaving the niche ruined, and people scared to invest in it again after the damage that was done.
I've attended every single Google I/O conference since the first one*, and for me personally this was the least compelling one ever.
It was pretty much all AI. I'm sure there are lots of other innovative things happening at Google that just didn't get any floortime. It felt like a couple of the sessions crammed "AI" into their title simply to be included.
(*obviously minus the ones that were canceled/restricted due to Covid)
31 comments
[ 20.7 ms ] story [ 1865 ms ] threadDo developers care any more what Google announces?
Isn't all the trust gone?
Every Google product announcement should come with a planned cancellation date.
The Google hate on HN also contributes to fewer posts or posts being flagged from greater visibility
The problem is that when you and three kooks say that Google is bad, people are generally not able to tell the difference between the kooks and you, even if what you are saying is quite sensible. Sigh.
Every business desperate for a boost in cash (and a pool of desperate LinkedIn posters) are pretending that they just invented machine learning in the last 18 months.
"AI" has become synonymous with either "snake oil" or "desperate marketing".
The government has sunk billions into EV incentives for instance, but consumers are currently struggling with inflation for essentials like food and shelter. A slightly cheaper EV isn't really a priority for many, which explains why EV inventories are piling up.
Simple test: Can a regular consumer derive value in less than 20 keystrokes that they already know how to type (eg English)? How about 50? 100?
LLMs: yes
Crypto: no
I suppose we’ll see in 2-3 years how many businesses maintain their newly found focus on “AI”. I’m betting that the organizations that aren’t employing internal teams training their own models will drop off quickly.
A lot of companies are just throwing together an interface to OpenAI and praying for the best. Long-term this won’t work. Users will eventually demand 100.0% data accuracy.
Eventually the novelty will wear off.
The paper that set off the LLM gold rush was written in 2017 by 8 Google researchers in 2017, none of whom still work at the company. Can you imagine where Google might be if they embraced this 7 years ago?
Instead, they're making "me too" announcements taht so reminds me of Microsoft with Bing or Meta with, well, just about anything these days (particularly IG and Tiktok "me too" features).
Google has other golden geese but almost all of them are acquisitions (eg Youtube, Maps, Android, Docs, Youtube). Chrome is the notable stand out.
Going through this list the first 53 or so are AI "me too" announcements. Google has blown a huge potential AI lead and has become an un-agile reactive behemoth, possibly even a dinosaur.
Embarrassing themselves 7 years ago, rather than now?
Products in that post: Gemini (Flash, Pro, Advanced), PaliGemma, Imagen, Gemma, Genkit, Gemini Live, Vertex AI, MusicFX, VideoFX, ImageFX, Veo, Music AI sandbox, project Astra… I’m sure I missed a few.
17 are not directly related to AI: 61-64, 66, 68-72, 75, 76, 85, 86, 88, 89 and 91.
5 look like they involve AI (would need to check): 43, 49, 56, 60 and 67.
78 are clearly related to AI somehow.
https://en.wikipedia.org/wiki/Google_Now
It was pretty much all AI. I'm sure there are lots of other innovative things happening at Google that just didn't get any floortime. It felt like a couple of the sessions crammed "AI" into their title simply to be included.
(*obviously minus the ones that were canceled/restricted due to Covid)