Your post is right to think about a future of "PCCs" versus AGI.
The last couple weeks of AI updates show two big trends:
* smaller models are increasingly capable – Meta's new Llama 3.3 70B model apparently matches the capabilities of their "old" 405B model
* models and agents are increasingly deeply contextual – Google dropped a demo of agents running directly within Chrome, with all the authentication benefits that entails, and also launched the ability to have a live screenshare with Gemini AI.
The latter point enables AI to operate alongside you more easily. The former allows these models to be deployed in smaller and more portable form factors.
The end result is a future in which AI is everywhere, all the time. The ramifications are astounding.
Thank yo Justin. I never liked the name AGI :-) But also whatever we end up defining that as, we do have a distinction between these systems and how we will interact with them. This is such an accurate statement "The end result is a future in which AI is everywhere, all the time. The ramifications are astounding."
I was testing Meta's Lama models, still not ideal for work but nevertheless a great example of where we are going. I am actually a huge fan of Google's systems - my students get tired of me saying how incredible they are:-)
It is. I haven't been following Google's developments very closely. 2025 is clearly going to be the year of agents. Is any of this available to the public yet?
Not in Europe! I think in other parts of the world it could be in Google AI Studio - follow Paige Bailey to see them in action - https://x.com/DynamicWebPaige
Progress has historically created more jobs than it destroyed, such is the natural progression of economic growth.
Attempts to regulate or “protect” jobs inevitably make people worse off.
To be fair, there is a chance that AI will be different. The pace of advancement may exceed society’s ability to adapt, or as I discussed here, if AI succeeds as a perfect replacement for human cognitive labor, then it could presumably destroy all “jobs” in the traditional sense and collapse the value of human labor.
I just don’t see any option but to muddle through. We can’t hit the “stop” or “pause” button, we must march forward.
Totally agree - as my research on robots showed the net effect is job gains.... whatever comes now, we will find a way to improve human life - - 100% "I just don’t see any option but to muddle through. We can’t hit the “stop” or “pause” button, we must march forward."
Your post is right to think about a future of "PCCs" versus AGI.
The last couple weeks of AI updates show two big trends:
* smaller models are increasingly capable – Meta's new Llama 3.3 70B model apparently matches the capabilities of their "old" 405B model
* models and agents are increasingly deeply contextual – Google dropped a demo of agents running directly within Chrome, with all the authentication benefits that entails, and also launched the ability to have a live screenshare with Gemini AI.
The latter point enables AI to operate alongside you more easily. The former allows these models to be deployed in smaller and more portable form factors.
The end result is a future in which AI is everywhere, all the time. The ramifications are astounding.
Thank yo Justin. I never liked the name AGI :-) But also whatever we end up defining that as, we do have a distinction between these systems and how we will interact with them. This is such an accurate statement "The end result is a future in which AI is everywhere, all the time. The ramifications are astounding."
I was testing Meta's Lama models, still not ideal for work but nevertheless a great example of where we are going. I am actually a huge fan of Google's systems - my students get tired of me saying how incredible they are:-)
It is. I haven't been following Google's developments very closely. 2025 is clearly going to be the year of agents. Is any of this available to the public yet?
A short vide on the Google agents - https://www.youtube.com/watch?v=Fs0t6SdODd8
Is it available for use or demo yet?
Not in Europe! I think in other parts of the world it could be in Google AI Studio - follow Paige Bailey to see them in action - https://x.com/DynamicWebPaige
The livestreaming functionality is available here: https://aistudio.google.com/live
Progress has historically created more jobs than it destroyed, such is the natural progression of economic growth.
Attempts to regulate or “protect” jobs inevitably make people worse off.
To be fair, there is a chance that AI will be different. The pace of advancement may exceed society’s ability to adapt, or as I discussed here, if AI succeeds as a perfect replacement for human cognitive labor, then it could presumably destroy all “jobs” in the traditional sense and collapse the value of human labor.
I just don’t see any option but to muddle through. We can’t hit the “stop” or “pause” button, we must march forward.
Totally agree - as my research on robots showed the net effect is job gains.... whatever comes now, we will find a way to improve human life - - 100% "I just don’t see any option but to muddle through. We can’t hit the “stop” or “pause” button, we must march forward."