Its pretty well known in #software development that some can be 100x more #productive than others. Google says using #AI internally makes them 10% more productive, that seems to be a safe estimate. So 0.1x.
Seems all these billions in investment are missing out on the real gains. Another data point supporting the case that AI companies are not focused on making us productive, but being good enough so management can get rid of pesky humans. More highly productive humans weakens the boss' power
I wonder if the way the #US and #Israel are leading the #IranWar has something to do with too much reliance on "yes men" #AI setup. There is all the news about #Anthropic #Claude being essential to their #defense planning. It seems they are moving tactically fast, killing key leaders, bombing so many things. But it looks strategically stupid. They are surprised that #Hormuz is closed. #Gulf states' bubble of perceived safety has popped, etc.
https://www.theatlantic.com/ideas/2026/03/pete-hegseth-strait-of-hormuz-iran/686368/
The #US military industrial complex is not even being intelligent about this war in #Iran. They are stuck in a bubble thinking that their overwhelming air power can do anything. This war has been in the making for years, so they had plenty of time to think about it.
"Iran has demonstrated it can escalate the costs of the war for #Washington far beyond its military capabilities to meaningfully counter the US-Israeli attack directly.
Putting the fundamental issues with GenAI aside for a second, I think I see a lot of parallels between self-driving cars and AI-generated programming. People are mostly okay drivers and okay programmers.
But a machine is always going to be faster and more precise in highly controlled enviroments. Except we don't live in highly controlled environments.
The political risk I see is that we start heading towards a world that's a highly controlled environment, a world built for LLMs not humans.
When designing #secure systems, understanding how people fit into the user experience is essential. Technical measures are only one part of the system. Without understanding the audience and their expectations, #security problems will surely arise. The more #mainstream the audience, the more important the #UX becomes in providing actual security. Security is a not a goal for users.
@uradotdesign just put up nice post on this topic: https://ura.design/en/blog/perfectly-useless
😵 🤯 Last year a #Meta Lobbyist became a new head of the #Irish Data Protection Commission (DPC). Today, the former head of the DPC announced that she is now working for MHC - the Irish law firm for Meta that was previously defending Meta before the #DPC in #GPDR cases - basically a 360° seat swap between the Irish regulator and Meta.
This is not just extreme #CorporateCapture of a regulator and makes #Ireland a laughing stock within the EU, but moves like this also undermine any public...
@Luanti Hey 👋 I'm a long time supporter of the project. I'm with F-Droid and in that context, we're putting together a funding proposal and we'd like to include Luanti, if there are contributors who are interested. Please get in touch! My matrix account is best, otherwise email or here on Mastodon.
A concrete example of how #AI saves #money for #PayPal: I have an old account with 150€. I want to get the money out but its locked. AI says I need to do more verification to prove I'm me. There are five pieces of info I had to provide. After months, four were marked verified. The fifth fails every time I try. Only #ChatBot support is available, which just points to me to docs. So they'll just keep my money knowing I won't sue over so little money. AI makes locking accounts like mine cheap.
A protected public forest in Kenya was being carved up behind closed doors; its trees cleared, its soil broken 😱
But, activists took the matter to court... AND WON!
A sword of Damocles over F-Droid https://youtube.com/shorts/CC0ihW8aNEM?feature=shared @eighthave #DMAVoxPopuli
Tim Sweeney signed away his right to criticize Google until 2032 https://www.theverge.com/news/889595/tim-sweeney-signed-away-his-right-to-criticize-google-until-2032
@eighthave couldn't resist the pun (Episode 1, recorded some time ago), and I couldn't resist making another short out of it! https://www.youtube.com/shorts/H7QoVdmzbWI #DMAVoxPopuli
"Expanding the number of developers in Google’s database will only make it a more frequent target of legal demands."
#Google #Android #AndroidDeveloperVerification #GooglePlay #privacy #legal #developer