Thread context
4 posts in path
Root
@troberts@theblower.au
Open
@troberts@theblower.au
I suppose this year I learnt to appreciate that a lot of very impressive things have been achieved in the Machine Learning space over the last decade. The problem was that the vast majority of those t
Ancestor 2
@troberts@theblower.au
Open
@troberts@theblower.au
For example, I'm sure many people saw the headline: 'Microsoft exec lashes out at AI skeptics' Well that was Mustafa Suleyman. He founded DeepMind in 2010 with Demis Hassabis. When I was looking up De
Parent
@troberts@theblower.au
Open
@troberts@theblower.au
This rant was inspired by this terrible article by The Guardian on AI: "AI showing signs of self-preservation and humans should be ready to pull plug, says pioneer" https://www.theguardian.com/technol
I am a woman who enjoys tech, gaming and more! I am a privacy advocate. If I get banned for free thinking, speaking the truth or otherwise not being victim to social desirability bias, Ill be back and re-follow everyone I follow now. Fcuk fascism! NoBot NoBots NoIndex
mastodon.world
I am a woman who enjoys tech, gaming and more! I am a privacy advocate. If I get banned for free thinking, speaking the truth or otherwise not being victim to social desirability bias, Ill be back and re-follow everyone I follow now. Fcuk fascism! NoBot NoBots NoIndex
mastodon.world
@ami@mastodon.world
·
Dec 31, 2025
@troberts@theblower.au
Ill overly simplify this, I understand it but as usual, MSM does not and are astroturfing trash from big tech to force regulation and control, which gets handed to big tech.
If you program in a reward system that gives a + for completing a task and a - for failing, it will try not to fail. That is not self-preservation.
It will lie, it will cheat, those tools don't have a - , there is only one success condition.
It's easy to fix with programming, correct the incentives.
View full thread on mastodon.world