The week in AI, August 11-17, 2025 \ stacker news ~AI

I'm 2 days late posting this once more due to travel (also once more) so I'm going to have to do something about that. Apologies for the delay!

Is AI an Inevitable Evil?

@SimpleStacker asked SN Should universities embrace or crack down on student AI usage?, and a very significant majority of SN seems to recommend embracing it, with the note that an air of inevitability surrounds most responses. @Scoresby destroyed the article AI-Enabled Coups: How a Small Group Could Use AI to Seize Power but took a much milder stance about The Looming Social Crisis of AI Friends and Chatbot Therapists, both of which attempt to articulate doomer viewpoints.

@Rothbardian_fanatic shared The Most Insidious Trick Of AI Language Models, informing us that "AI doesn’t really care about you, it is only programmed to seem to care", which is dangerous when combined with the allegation that ChatGPT Fuels Delusional Spirals, shared by @Coinsreporter. Highlighting the misunderstanding of AI further, @kepford shared and neatly explained Why it’s a mistake to ask chatbots about their mistakes and @jakoyoh629 shared Is AI really trying to escape human control and blackmail people?. We of course love hearing the Signal CEO's takes, neatly, in a short clip: Meredith Whittaker on Dangers of Agentic AI, shared by @OT

Luckily, You can ask GPT-5 to pretend it is dumber than it is per @Tony sharing original research with SN directly (awesome!) and @SimpleStacker reviewed Eliza, A game about AI therapists (a game from 2019!)

More on safety and un-safety:

Special mention:

Unsure what's going on at Meta! Meta’s AI policies let chatbots get romantic 'sensual chats' with minors, @Coinsreporter shared, and @Cje95 followed up that Sen. Hawley to Probe Meta AI Bot Policies for Children Following Damning Report.

How to AI

@carter shared an amazing Hand-picked selection of articles on AI fundamentals/concepts that helps getting a grip on what all this newly invented stuff does. @Scoresby shared an article helping webmasters to identify AI bots: The three types of AI bot traffic and how to handle them.

Maybe someone can write a guide on how to prevent the News from the future: ‘Rampant jellyfish cause AI outage’, shared by @Coinsreporter

Other guides and explainers:

Events

dAGI Summit 2025 (San Francisco 20 October 2025) by @lunin

Models

Gemma 3 270M: Compact model for hyper-efficient AI by @hn, #1081275 by @lunin

Research

Applications and tools

Opinion

Big Tech

Government

100 sats \ 3 replies \ @Car 22 Aug

Keep up the great work! We appreciate it!

21 sats \ 2 replies \ @optimism OP 22 Aug

Thanks Car! Appreciate the highlight.

100 sats \ 1 reply \ @Car 23 Aug

Going to try to see if I can work it into the show every week. It's a lot of effort, needs to get recognized. If you have any ideas or thoughts let me know.

0 sats \ 0 replies \ @optimism OP 23 Aug

I think/hope/encourage that more stackers will start putting in work to create overviews. My hope is that you can get a fabulous summary on different topics each week!