pull down to refresh

Everyone now understands the value of AI as a tool for code generation, text translation, summarization, photo generation, and a handful of other simple digital tasks.
But what are the AI use cases that might be possible in the future, yet aren't feasible today for one reason or another?
I'm specifically interested in things that require you to squint out into the future and believe in something that isn't a consensus belief today.
Maybe it's because I don't see AI as a new thing in a material sense, because if you distill it down and it's just another algorithm... but I think eventually we're realize what we're actually using it for already... an interface to other algorithms.
We see it in agents and chatbots now, things like MCP, where natural language calls upon external API's in tools which then have their own algorithm or process.
This ability to chain functions from disparate sources makes it a new interface type for applications that are just a representation of many functions.
There's a lot of things we think of in terms of UI now that will probably just become MCP-adjacent tools in a language interface vs. a visual one, image gen vs. Photoshop for example. It's already consolidating time away from the browser (RIP keyword searches and stackoverflow), and voice based bots can obviate screens altogether in many cases.
We're probably less than a generation away from kids not knowing wtf a keyboard and mouse is. A pocket projector that can leverage your eye/hand/body movements, or glasses, will change how much we think about interfacing as the iphone did.
reply
110 sats \ 1 reply \ @JesseJames 5h
Firefighters, have them control strong humanoid robots with 100+ censors, so they (the firefighters) can do their job more safely.
reply
0 sats \ 0 replies \ @kr OP 5h
Good idea, could also probably work for police officers
reply
100 sats \ 3 replies \ @kr OP 6h
My first thought was Tesla's Optimus robots using AI models and vision to perform more complex human labor tasks, but even that is flirting with the "already a consensus belief" line.
reply
11 sats \ 2 replies \ @optimism 6h
I liked #1218570 this morning
reply
100 sats \ 1 reply \ @kr OP 6h
Nice, on the surface it looks sort of similar to the Boston Dynamics/Tesla/Figure humanoids, is there a unique angle that this product is taking? Or a new set of use cases that they are aiming to solve?
reply
100 sats \ 0 replies \ @optimism 6h
It's open source, also the hardware. BOM
reply
To choose the Prime Minister of Nepal hahahaha #1219095
reply
221 sats \ 4 replies \ @k00b 6h
Intelligent, autonomous, self-repairing infrastructure: #1210423
reply
0 sats \ 3 replies \ @kr OP 6h
So AI will be used to make sure all the sensors are always lined up correctly?
reply
10 sats \ 2 replies \ @k00b 5h
yeah but you can generalize it:
  1. infrastructure that doesn't exist because the environment is dynamic, can exist
  2. infrastructure that's high maintenance and expensive because the environment is corrosive, becomes low maintenance and cheap
reply
0 sats \ 1 reply \ @kr OP 5h
When you say infrastructure, what other places could this kind of system work?
Can this solve for plumbing, electricity, sewage, etc...?
reply
20 sats \ 0 replies \ @k00b 5h
Yes, I think when you make arbitrary things intelligent and have some kind of body the intelligence can control, it allows for a a lot more solutions. e.g. the location of a solar panel.
reply