I’m growing increasingly terrified. And I don’t think I’m being melodramatic about it. This is required reading.
#ai #microsoft #bing
https://open.substack.com/pub/tedgioia/p/over-the-course-72-hours-microsofts
Over the Course of 72 Hours, Microsoft's AI Goes on a Rampage

I thought the AI story was bizarre last week, but that was nothing compared to this

The Honest Broker

My intent was misunderstood because I didn’t take the time to explain my fears. Let me be clear:

I’m not scared of the current state of #AI bots like #ChatGPT or the #Bing incarnation. I’m scared about the powerful profit-driven mega-corporations having any control over their future and the influence they will have over a population.

Not sentient now, no. Still capable of causing real harm even in their infantile state? Yes. What does 2030 look like, then?

@killyourfm I currently see three groups of "AI" tools:

- the ones we hear about on the news and in the media. Mainly showcases or technical demos.

- the emerging services or those already in tools we use. Think M365 Editor, Google Assistant, GH Copilot.

- the "industrial" tools, those that we do not see but automate server actions or manage bulk data.

I do not think that AI is a bad thing, but that controls and codes of practice are needed - possibly similar to open source licenses.

@killyourfm you nailed it. Nothing else to add

@killyourfm
In that regard: AI will consume lots of energy for training the models. Microsoft is putting PV on their data centres and offices.

2030 will already be post the infliction point for climate crisis. Expect raising sea levels, droughts and conflicts that will arise. Perhaps not exactly in 2030 but in the decades that follow.

Now, will it be our priority to be entertained online?
Perhaps not as much. But Microsoft is placed in many state offices.
#bing #chatgpt #ai

@killyourfm That being said, those corporations will only have as much as power as society is willing to grant them.

I'm sure many people won't be considering switching. But look at what happened in Turkey a couple years ago when the government blocked Wikipedia (people turned to TOR browser).

So it's on us technical folks to prepare the infrastructure to offer alternatives when they are needed. Federation will play a big part in it.

#bing #chatgpt #ai

@RyunoKi This deserves a boost. Those are some wise words.
@killyourfm Well, there goes my ability to sleep soundly at night....
@killyourfm Nope. I'm still skeptical. This is a robot that's been trained on human language. Human's write a lot of weird things. This is not sentient, this is a simulator of that and it's doing it's job.
@killyourfm eh, GPT is just a language model, chatGPT tosses in a live search index to the mix and uses a mix of rewriting and re-seeding information to carry a "conversation", like the 2022 thing is just because it has no concept of actual time, it's just repeating that because all the search results were talking 2022. The only scary thing about chatGPT is the thought that people will use the information it provides as "truth". One day the AI itself will be scary but this ain't it.
@killyourfm Why? It's not sentience, it's madlibs. It's just producing statistically likely real-sounding strings of words.

@Magess
@killyourfm Out of curiosity:

Have you heard of ELIZA before?

@RyunoKi @killyourfm I felt like the answer was yes but still had to look it up. I know I played with it back in the dark ages but couldn't tell you how I came across it.
@killyourfm Remember Tay?
"Tay was an artificial intelligence chatbot that was originally released by Microsoft Corporation"
https://en.wikipedia.org/wiki/Tay_(bot)#Suspension
Tay (bot) - Wikipedia

@testman Sure do! An inevitable end when trained by language on the internet.
@killyourfm this shit goes harder than expected
@killyourfm Also, what happens when all the places they are stealing their content from get tired of it or go out of business? How do they think this is sustainable in the long term?