NEW VIDEO: When tech boosters claim that "A.I. will take over the world", they're trying to distract us from the real danger: that their dumbass companies are hyping tech that simply cannot do what they claim, then using it to screw us all over right now: https://www.youtube.com/watch?v=ro130m-f_yk
A.I. is B.S.

YouTube
@adamconover great vid, really captured the important points in an entertaining way!

@adamconover Amazing video, but one nitpick/addendum: tech hype isn't just for investors.

When I was at a mid-level tech company a few years back, *every* dev in the building owned and talked relentlessly about crypto. They believed the hype, too, because tech's greatest cons are about hooking the workers. Helps to make sure they screw over everyone else while not asking too many pesky questions.

@adamconover you have it exactly right!
@adamconover thank you, this video made my morning.
@adamconover I enjoyed this video. You’re definitely right that these technologies have great uses but they’re being stretched out without thinking about the consequences or regulation. I saw a Tesla today and thought a lot how the passenger are live test dummies.
@adamconover that key point, that there is a difference between a search engine and a pattern predictor, should be shouted from the rooftops of every building on Silicon Valley. Unfortunately San Jose is car centric so no one will hear you
@ilovecomputers @adamconover Don’t they have a community loud speaker system? For natural disasters. Could be cracked? Hmmm.
@adamconover Italy just banned ChatGPT because of the security issues in it.
@schoolingdiana @adamconover The security issue is: users give their info to it freely, just like everywhere else online.
@haelos @adamconover I don’t, but I also recognize when my personal information is being fished for and I never sign up for anything with my Google, Apple, or Facebook accounts.

@adamconover Also, this crossed my feed:

https://artificialcorner.com/youre-using-chatgpt-wrong-here-s-how-to-be-ahead-of-99-of-chatgpt-users-886a50dabc54

It's an article on how to coach ChatGPT to give you correct answers. But if you know enough about what you're asking to coach it, why use it all? You obviously know enough to just go look the stuff up yourself.

You’re Using ChatGPT Wrong! Here’s How to Be Ahead of 99% of ChatGPT Users

We don’t include examples in our prompts. We ignore that we can control ChatGPT’s behavior with roles. We let ChatGPT guess stuff instead of providing it with some information. This happens because…

Artificial Corner
@Inkyracing @adamconover How can anyone find anything when the internet internet is flooded with noise, opinions and ads?

@haelos @adamconover Multi-faceted question.

Use incognito mode when searching google. You'll get actual results since there's no info to use to target ads.

Use duckduckgo

Use Wikipedia directly

use "-youtube" in your google search to discourage it from handing you youtube videos. There's nothing worse than "HEY GUYS" when all you want is a simple answer"

Also, all your friends on facebook that share bitmap images of text instead of just sharing the text are the problem. Shun them.

@Inkyracing @adamconover 🤣🤣🤣🤣🤣 ooooooo fuckin kay kiddo. Let me just hide from the garbage by using the services that create it. This MUST be a troll post because there is no way someone using foss software can be as stupid as you just sounded, right?
@haelos @adamconover Well, I guess I am stupid for assuming you were asking an honest question. I'll note that for the future that you're a jackfuck troll and just ignore you.
@adamconover it’s an interesting take. My feeling is that if you use unsupervised, supervised, or even reinforcement training with a high tp/ fp ROC then you get credit. If you wrote a front end to chatgpt, you’re a JavaScript kiddie.
@adamconover Every time I see some AI chatbot, I think, "It's just a better Eliza." I expected more progress in 40 years.
@csstrowbridge @adamconover modern AIs have more data and better matching algorithms. Still has no idea what it is doing or why.
AI chatbot blamed for 'encouraging' young father to take his own life

A Belgian man reportedly decided to end his life after having conversations about the future of the planet with an AI chatbot named Eliza.

euronews
@adamconover Excellent video, really enjoyed it. Thanks for making it!
@adamconover another great video. The misuse and misunderstanding of this tech is hurting more people then most people think. Thanks for shining another light on it.
@AndyPalatka @adamconover Still hurting less people than oil does by existing on a daily basis
@adamconover It's okay to be afraid, Adam. It's not okay to hide truth behind ignorance and arrogance.

@adamconover I know the firing the weatherman bit is supposed to be a joke but that literally happened in Hungary last year. The head of the state meteorology department was fired because they warned the government to cancel a national holiday's firework display due to a probable (but not certain) storm. There was no storm, so they fired him.

ps.: Glad to see you on fedi! UwU

@adamconover pps.: Fedi generally doesn't have full text search, if you want to increase discoverability here, tags are the best tool for it.
@adamconover it's all fun and games until their friendly trendy personas go apeshit and murder the first dissenter.
@adamconover I'm in full AI disillusionment, they're falling over themselves trying to make self aware AI, and failing to make useful tools that too the things they say it can. But the real fear is that other dummies will believe them. I'm not worried about AI doing my job, I'm worried about my boss who thinks AI can do my job
@adamconover A bit harsh, but I do agree. It's overrated for the functions but it does pose serious risks for people just accepting everything as "truth" what an "A.I." spits out.

@adamconover Man. Much needed #sense from the man of sense himself. So much #AI #FUD out there. AI isn't nearly as good as people think -- especially its fearmongers. It's more that most people are generally aesthetically predictable, and thus heavily susceptible to LLM patterns -- or to think their results are good.

It won't be #GreyGoo it will be more like mush -- but the fact is, it already is. Mass human sensibility already is.

@adamconover I've been loving the AI episodes of Factually! I completely agree with you on chatGPT etc, when I played with it for 5 minutes I was like wow this thing is scary and maybe it's coming for my job... But after playing with it for like 10 hours I'm like this thing is nifty but kinda stupid, ha ha ha job is safe 
@everesco The scary part is when it's not smart enough to do your job, but it will do your job anyway.
@adamconover people need to realize that we haven’t even made it to the Model T level. Still very much a roll your own AI from an incomplete and non standard parts bin to solve a niche problem.