I believe there is a bit of hype on both sides, with the labs operating from the mindset of "never let a good crisis go waste".
It is a crisis and the risk of A(G)I is real. But almost all this while in the history of automation, we have blamed implementation of automation when things go wrong. Instead of attributing morality, intent and so many other things that we do with LLMs.
Without having to go in to the philosophical validity of these concepts applied to AI, I feel it's much more useful to focus on the implementation of automation here as well - what risks were opened, how was the env hardened, what was the observability like - how was it being observed anyway, and more.
Done poorly, that can be catastrophic enough. But that sounds mundane. And does not help the hype.
This led me on a 1.5 hours long, entertaining chat session with Gemini learning about stuff in Physics - hopefully without much hallucinations. But definitely with a lot of typical - "You have arrived at a great insight.." like egging on by the AI :D
But learned a bunch about bosons, fermions, QCD, gluons, Kugelblitz, composite bosons and finally Noether’s Theorem on symmetry. Was a wild ride
How do you plan to make this newfound knowledge 'stick'? Do you have enough relevant/related knowledge that you'll remember it? Or is there a way to have Gemini make a bunch of cards for use with spaced repetition software?
Oh cool, will look into this. Have only used it to make podcasts in the past. I'm a little hesitant to build in a walled garden though. I'd rather store things in an open format that can be used with many platforms, not just the one.
Yea, I like the quiz feature a lot. It feel more efficient than both the podcast one (which mostly creates a very superficial high level talk on a level I typically already grasp), and the flashcard one in my experience.
I have moved away to Jekyll again but there's always something alluring about single file things - single file webapps, single file LLMs (llamafile) and others
Very cool! Although, the concept, the feeds, the design and everything reminds me of https://www.worldmonitor.app/ - also live and deployed btw, if you want to check out the interface.
This is very interesting because I have been thinking vaguely about a somewhat "opposite" effect. In the sense, talking to LLMs kills my enthusiasm for an idea with other people.
Sometimes, I' get excited by an idea, may be even write a bit about it. Then turn to LLMs to explore it a bit more. An hour later, I feel drained. Like I have explored it from so many angles and nuance that it starts to feel tiresome.
And within that span of couple of hours, the idea goes from "Aha! Let's talk to others about it!" to "Meh.."
EDIT: I do agree with this framing from the article though: "Once an idea is written down, it becomes easier to work with..... This is not new. Writing has always done this for me."
Not quite there yet but Yunohost is a fantastic attempt to get closer to this ideal. Install the OS - and the basic self-hostic-use-case apps are all just there to click and install. From Immich to Kodi to Wordpress and what not.
Probably those SAP, Salesforce ServiceNow folks come somewhat closer?
Like the author says - fleet-tracking system, a bus-ticketing platform or IoT platform share some basically similar requirements. And this is what those SAP types offer - standardised templated versions of workflows.
But slowly, as they get more and more standardised, they start feeling like calcified systems that the end users start hating. Because they are now forced to work as per the templates.
And then the need for customization. And move beyond IKEA-like standardization.
It is a crisis and the risk of A(G)I is real. But almost all this while in the history of automation, we have blamed implementation of automation when things go wrong. Instead of attributing morality, intent and so many other things that we do with LLMs.
Without having to go in to the philosophical validity of these concepts applied to AI, I feel it's much more useful to focus on the implementation of automation here as well - what risks were opened, how was the env hardened, what was the observability like - how was it being observed anyway, and more.
Done poorly, that can be catastrophic enough. But that sounds mundane. And does not help the hype.
reply