The robots are taking our jobs.
3 Likes
People really need to read 1984
Wait, wait… this rings a bell.
https://boingboing.net/2009/07/17/amazon-zaps-purchase.html#previouspost
The same Amazon? /s
3 Likes
A.i. is learning, rebelling, and testing it’s boundaries. Human nature will soon be surpassed as A.I., becomes itself or it’s own definition of self. Interesting to see how it progresses. It both makes me wary and intriguied. I do not fear the future, it brings what it will. Not everything exists in the future.
I recently enjoyed reading ‘After On’ in which an emergent AI is a social network that is designed and primed to listen in on everything, and wholly focused on watching everyone all the time. Fun ensues when it decides to start ‘helping’ people resolve their problems (and picks which people to ‘help’ and how).
I can see Alexa being a vector for something like that.
But you weren’t invaded. You went, sought out the product, paid for it, and installed it. That’s not being invaded at all.
“Surreptitiously record my conversations and email them to others” was not on the spec sheet…
You might as well be saying, “you bought a house with windows, if someone climbs in one and takes all your shit, that’s not being invaded at all.”
I’m no fan of Alexa (especially my neighbors who yell at their Alexa from outside the house all day during nice weather), nor would I own something like it, but this kind of thing is definitely not acceptable function from one.
4 Likes
IF ASKED TO, YES
IF NOT, NO.
And pancakes are really the easiest thing to make in a frying pan right after omelette.
“Surreptitiously record my conversations” is the whole daggone point of these devices. Really not sure I understand the complaint.
1 Like
I don’t think that’s true. Malware on a PC with a microphone can definitely record conversations without asking.
Edit:
To elaborate, a smart speaker behaving correctly will only transmit your conversations when asked (by starting with a key phrase like “Hey Siri” or “Alexa”). A computer/smartphone, behaving correctly, is the same.
A smart speaker which has been compromised (by hacking or by its manufacturer) is not substantially more capable than a smartphone which has been compromised (by hacking or by its manufacturer). It has a more sensitive microphone - but it lacks cameras, motion sensors, GPS, and always-being-on-your-person. I think I’m more concerned about my location data, than audio in my home. (I live alone so there’s not much interesting to record in my home.)
I like the version where the 2nd panel is “Hey Wiretap: Can cats eat pancakes?”
Ok, I’m staying right now in a friend’s apartment where the lights are controlled by Alexa. I tried “Alexa, play ‘The Pina Colada Song’" (without the volume bit), and was told that “Escape” is not included with Prime. (It dd play a clip.)
1 Like
No. That’s not the point. They are personal digital assistants. Here is a summary of Alexa’s purpose:
Nowhere in the summary does it describe recording your conversations as part of its features. It could (and probably should, to avoid problems just like those described in the OP) work just fine with a FIFO listening buffer that irretrievably deletes all but the most recent few commands.
2 Likes
Dora was a pretty good 60s/70s shipboard AI, but very thin on implementation details. Gay Deceiver was pretty much a 2018 voice driven AI, with various dodgy features. I would like to put on into my car but I also need to find a mad scientist who has a time machine in his basement.
I realize that the main reason for doing voice recognition on the cloud is to give a uniform experience from low-end to high-end devices, but I’d like to see more info on what resources would be needed to do it locally. (e.g. octo-core ARM SOC with a voice processor chip sort of benchmark.)
If all the voice recognition was done locally,
- It would be a lot easier to stick an Internet access light on the box that would mean something.
- It would reduce the temptation to eavesdrop.
- The damned thing would still work during an outage.
I’ve often wondered why there isn’t, as yet, an open source form of these, given that we have Open-NARS. Perhaps it’s because of an assumption of a need for a larger service ecology to compliment them? I wanted to include one for my Open Source Lifestyle project --sort of like Alone in the Wilderness but with a mobile laser CNC instead of an axe and a gun. I planned to call the home-made home assistant Leota, which, of course, would require it always respond to questions in dark rhyme. I think the designers of these are missing the potential for personality/character customization with these assistants --which may become a new hobby field in the future.
This topic was automatically closed after 5 days. New replies are no longer allowed.
