Is it possible with today's tech that we each could have our own Jarvis? Not vision just a Jarvis. Like being able to tell your house to dim lights when coming home, or maybe start the crock pot?

Pretty sure Google home can do this, it can control and automate Phillips hue lights, and there are compatible smart plugs, that you can plug your crock pot into and it can turn it on with voice commands

No thanks. I don't want to be the chair people at the beginning of WALL-E.

!homeassistant.world

https://www.home-assistant.io/

Im sure it has some integrations for this

I know a guy who automates his home as a hobby. It's disturbing to see a person who has read 1984 and is aware of the weaponization of US Tech, answer my question of "Why?" with a response of the convenience while he argues with Alexa in the dark over which lights he wants on.

Before leaving, I told Alexa "set a wake up alarm in the master bedroom for 4am and play Rebecca Black's Friday at maximum volume while flickering the lights on and off for 10 minutes. Repeat this alarm on "his birthday".

I already got two nasty texts from this guy. He is still doing it.

To be fair he doesn't have to use Alexa (or Google or any of the tech firm's BS electronics). What they do is provide an easy avenue for those who aren't tech savvy, or are lazy to incorporate smart home features into their homes without a lot of extra fuss. Given how these products seem to degrade over time it might be better to just suggest he look into home assistant rather than sabataging him, but your mileage may vary.

Is it possible? In many ways, yes. Combining a cutting-edge LLM with something like OpenClaw and lots of IoT (Internet of Things) connectivity, it could do a lot of what Jarvis does. At least in managing a building and doing some chores for a human locally.

Is it a good idea? Arguably hell no! LLMs are remarkable, but they have blind spots. Notably with being exploited and gamed by malicious players. Or just in making strange decisions or hallucinating shit. And there have already been a lot of horror stories about OpenClaw-enabled LLMs either getting exploited or just unilaterally doing detrimental things under (mostly) good intentions, and leaking PII or bank details or similar. Then you factor in all the security holes that IoT devices can be prone to, and the expenses of buying all the needed technology, and paying for all those LLM API tokens. Not to mention the environmental concerns of using a very powerful cloud-based LLM. I'm probably forgetting some other negatives too.

ETA: If you're going to try creating a Jarvis type system for home automation and sooner online tasks, really make sure you take security and financial precautions. Don't give it access to bank accounts and sensitive PII, ensure limits are in place for token usage, be very specific with any prompts you use, and for the love of God lock all devices and systems down behind really strong passwords/passphrases and (where possible) MFA that goes directly and only to you on a separate device or number.

And just like Jarvis, he'll have access to that commenters personal data, and so much more! Fuck Google. Fuck a lot of them but fuck Google in particular.

Have an AI walk you thru setting up Home Assistant, which is often considered the best software for doing this (and it's probably free). What you want is totally doable and done by many people as a little tech hobby and helpful upgrade to the home stuff. It's not super hard, but having an AI be your mentor will make it much easier. Project cost might be a few hundred dollars, since you will want a cheap mini computer and a few electronic bits to allow communication between various devices. Of course, the individual devices e.g. nice phillips hue lights, are all additional costs.

You could also check out the HomeAssistant subreddit, but it's often a bit technical and not focused on someone starting out like you are.

Totally possible. Thequestion is whether it's worth it. Science fiction is cool because of the fiction, not because of the science.

When it becomes science reality it instantly loses it's luster. Because science fiction doesn't tell you about the things that suck about the tech.

Has there ever been a science "fiction" book that actually dived in to the mechanics just to explain what's going on and years later added on to those mechanics and it became fact? Generally curious.

Depends on what exactly you mean.

Star Trek famously took scientific concepts that were in early development at that time and finctionalized them. Some of them then were developed in reality. They didn't "invent" them, but they did popularize them.

For example, the first early prototype work on touch screen was published in 1965, and Star Trek introduced them in 1966. At that time the concept of touch screens was not widely known in the general public. Touch screens did become a wide-spread product much later.

Research on speech recognition started in 1960, computer-based speech synthesis in 1950 and Chatbots in 1964. Neither of them were any good in 1966 when Star Trek used the concepts to create a "computer" that one can talk to. They neither invented the components of that, nor did they invent the combination of all that. But when they used that in fiction, reality wasn't nearly ready to actually deliver on these promises.

In general, good science fiction usually uses stuff that is right now in research. Bad science fiction makes shit up.

yes