The Short Story of AI

This last week was … sort of odd on a personal level.

Got some more work done on Gregory AI, making the system able to monitor different research topics for different teams.

And lately have been looking at #Zotero, which in my view is much more useful and powerful than Mendeley. My goal is to figure out how to sync it with my cloud drive and access it on all devices. If enough people show interest I may sort out my bookmarks for zotero and write up how to use it and the plugins.

The operational tasks that took over my schedule pushed away the time I like to have for planning and scanning what’s happening with each project. Getting back to that mode released more endorphins than I’d like to admit.

The AI story so far

Next time there is a big breakthrough in technology I need to remember to be more methodic in keeping track of it.

In a quick retrospective:

OpenAI was a huge hit, and big players tried to slow down AI research because it was a “risk for society”.

We discussed the impact of AI to our work. Then we questioned the ethics of training it with a skewed values system and adding guardrails to prevent misuse.

This stand lost its footing when a Times investigation revealed that major technology companies were bending and breaking rules in their rush to develop new artificial intelligence systems.

Later the discussion was about the right to use copyrighted work to train the image and language models. Fast forward to today and big tech is lobbying for “fair use” status to train AI systems.

Right now, we are still trying to find AI’s promised productivity boost, while big tech lays off employees.

This last week, Apple announced the new AI tools that fit seamlessly into their devices.

On previous notes, I wrote about the need to make AI ubiquitous to really make it useful, and Apple is following that route.

All this is bringing back the discussion about how much hardware and energy we need to power this new wave of technology.

And in the race to find training data, Meta is turning to Instagram and others are selling User Generated Content.

Making its move, Adobe is trying to train AI models using the files their customers stored in their Creative Cloud. Some people were even locked out from using the Adobe Suite because the new terms and conditions violated their company or customer policies.

There were also issues with Microsoft’s new Recall feature, but that’s not something I followed closely.

All of this feels like every player is scrambling and nobody has a plan.

Your TikTok videos for this week

How on earth did corporations come to have more rights than you? Watch this.

Is anyone else bored of just digitally streaming music? Are cassettes making a comeback?