@JessicaBRiedl gets it right with "authoritarian nations moved there gradually yet persistently - the govt commits some outrages, waits for the public to get used to it and stop complaining, and then pushes for the next round." So hard to keep up.
- Comey indicted for tweeting a number.
- Trump FCC threatens ABC's broadcast license.
- Trump defacing more govt institutions with his name and picture.
- Trump's kids cashing in on huge govt contracts.
I'm always torn on whether to speak up on the daily drumbeat of Trump's creepy, corrupt, Putin-style authoritarianism.
On the one hand, this has become so standard under Trump, and everyone's opinions on him have long hardened, so why bother? Just keep your head down, post on economic policy (what happened to Econ Twitter?), and don't be distracted by the daily news cycle outrages. That's the standard position in DC policy circles.
On the other hand, isn't this defeatist passivity how authoritarianism thrives? Government actions that would have induced a nationwide backlash 15 years ago are now just the day's background noise. So the bar moves.
It’s a familiar path: Most authoritarian nations moved there gradually yet persistently - the govt commits some outrages, waits for the public to get used to it and stop complaining, and then pushes for the next round. Authoritarian governments count on the population having a short attention span while they steadily boil the frog. But when you think about how far Trump has pushed just in the last 15 months, it’s fair to wonder where we'll be after the next 33 months of his term.
So I’ve been torn - speaking up a lot during the ICE protests, then just focusing on econ for many weeks. I don’t want to bore people who’ve all made up their minds on what is happening in this country - and I've been trying to spend less time on this toxic hellsite for my own mental health.
But I also think us trying to be all cool and above-the-fray is exactly what gives the administration the green light to once again push harder towards corrupt, statist, personality-cult Putinism. I don’t really have an answer here. Although this being Twitter, I’m sure that any possible approach (whether passive, hair-on-fire, or in-between) will inevitably get me yelled at here.
@emollick Why do I now picture @clairevo spinning up a creative agent in her little Claw empire to do just this? "Your nemesis is <insert fave AI poster>... Always respond with words like 'insouciant' and 'bete noire'. On Tuesday, speak like a pirate. Drunk on rum, but good rum."
Perhaps my proudest social media achievement of 2026 is having no idea who Clavicular or Hasan Piker are or why they keep showing up as trending.
Replies closed to keep anyone from shattering my blissful ignorance. 🙃
My openclaw twitter mention block cron job is working unreasonably well. Turns out AI is really good at detecting spam/reply guy/promo stuff.
Runs every 5 min and cleans up my mentions - I actually see useful replies now and Twitter got pleasant again!
@clairevo Stupid European cruise starts on April 17. GRRRR. Oh well, life is full of choices. LOL Alternative medium easy mode incoming.
(and if my hubby is reading this, the "stupid" was being humorous) 😉
@MicahBerkley@Kling_ai Seriously. Every cat owner knows this should end with Lisa casually strolling in from another room -- that you KNOW you checked thoroughly, giving a slight yawn and stretch. /where the hell do they hide?!?/
@SvetaSchnee That sounds very frustrating. Probably won't help right now, but we never know what is going on in people's personal lives outside of WoW, or work, or church, or even our neighborhood. Try not to take it too personally (I know, easier said than done).
@ChrisSBronkhors@JeffNylen@gmiller@HansMahncke@chipro addresses this nicely in her 'AI Engineering' book. If you hold all sampling variables fixed (e.g. top-p, top-k, temperature), the model should produce consistent output. However, the hardware used for inference can impact the output. Sampling makes them probabilistic.
@JYuter@ariehkovler Wait, what just happened here? Two people engaged in a Twitter debate and it didn't end in name calling? Maybe I stumbled onto the wrong website.
Nice job, gentlemen. You just made my day.
I was fortunate to hear Jake Mannix present at OAI's virtual conference early this year, so I was stoked to see that he will be presenting at the live event. I am sure his session will be can't-miss for those of us building enterprise-level agents.
Jake Mannix — Technical Fellow at @Walmart Global Tech — is joining Optimized AI Conference 2026.
His talk breaks down why today’s agent architectures fail at scale — and how to build governed, reusable, enterprise-ready AI agents.
If you’re shipping agents in production, don’t miss this.
Register to get 50% Off Tickets till Dec 26: https://t.co/DKfHGvyUrZ
Volunteer: https://t.co/4KNlXFCLKP
Join us on Discord: https://t.co/g9KNoIMIoV
⚡️#OptimizedAIConference #AIConference2026 #AgenticAI #AIEngineering #EnterpriseAI