I recently rewatched "The Lawnmower Man" [0] and was not disappointed. The vast majority of the comments I see promoting the notion of "AI" achieving AGI sound like the Jobe Smith character from the movie.
LLM/"reasoning" models will not manifest "AGI" anytime soon, as it would take 75% of our galaxy energy output to bring the error rate to average human levels.
However, several Neuromorphic computing projects look a lot more viable, and may bring the energy consumption needs down by several orders of magnitude. =3
> there is no way to confirm that this code even runs anywhere ever
I'm confused what this has to do with "open source" or how it affects public perception.
I agree with you that it's totally possible to lie about what is actually running in production and that sharing some code doesn't mean it's that code, but how is this a new problem?
The interesting part to me is how this anti-feature could become the primary source of value for AI if only it was easier for everyone to run and train locally from a blank slate and without the clumsy natural language interface.
Take the example of music. Most musicians probably don't want crap like Suno. What they actually want is a fountain of ideas to riff on based on a locally trained AI where they have finer-grained control over the parameters and attention. Instead of "telling" the AI "more this and less that" would it not make more sense to surface facets of the training data in a more transparent and comprehensible way, and provide a slider control or ability to completely eliminate certain influences? I'm aware that's probably a tall order, but it's what's necessary.
Instead of producing delusions left to random chance and uncurated training data, we should be trying to guide AI towards clarity with the user in full control. The local training by the user effectively becomes a mirror of that user's artistic vision. It would be unique and not "owned" by anyone else.
My experience with marketing pages is that they usually have a ton of inconsistent design requirements and change frequently.
Most "frameworks" are the wrong tool because they assume that the markup and design (HTML/CSS) won't change as much as the functionality (JS) when it's exactly the opposite situation.
All the consistency needs to be concentrated in the JS without the baggage of any particular HTML/CSS in mind.
The only aspects of a framework you should want are a flexible way to register event listeners onto the elements, and organizing the styles and callbacks.
In practice, this ends up looking like a static HTML file that is not up to the developers how to organize apart from the usage of CSS classes because it will be audited by many non-devs, a Sass build derived from design guidelines with some alt classes to contain the mess when people change their minds, and some very robust JS that you're gonna have to write almost entirely from scratch.
I still don't get why this scares some people off though. You won't ever need to (re)write that much JS. Every new page design is just remapping existing functionality to the new HTML IDs, and maybe every now and then adding new functionality. Most of your time will be spent in CSS which just plain makes sense!
I don't think anyone's job is copying "know-how". Knowing how goes a lot deeper than writing the code.
Especially in web, boilerplate/starters/generators that do exactly what you want with little to no code or familiarity has been the norm for at least a decade. This is the lifeblood of repos like npm.
What we have is better search for all this code and documentation that was already freely available and ready to go.
Do we accuse AI of having poor soft skills when it doesn't do what we want?
I think the problem is knowing how to bridge knowledge gaps. That just comes from experience and there are no shortcuts on either end of the gap.
Empathy does matter a little bit, but to focus so much on it is plain neurotic. Consider how much less friction there is when the interactions can be kept brief. Everyone is already familiar with the various situations and problems that can arise (like on a sports team). That's pure hard skills, not soft skills.
Posts like this are flamebait for the extreme ends of these gaps: stubborn mediocre programmers and arrogant dumb management.
On the flip side, if a user expects too much from a fitness tracker it can lead to unreasonable health anxieties.
A user trying to determine an accurate heart rate or blood oxygen level during exercise (not at rest) will find that the guidelines are too broad and the tracker data is too slow and noisy to get the feedback they want. They can get a rough idea of how hard they exercised and for how long, but a fitness tracker isn't necessary just for that.
"Jazz police are looking through my folders. Jazz police are talking to my niece. Jazz police have got their final orders. Jazzer, drop your axe, it's jazz police!"
reply