Hallmarc

Of course. I wouldn't expect logic to work. They have effectively become a semi-private paramilitary force that's full of untrained but deadly cosplayers.

There is of course a replacement rate below which humanity's numbers enter irreversible decline but even with climate change and other threats, it would take centuries. What you hear now is mostly whites trying to stop the entire world from turning... not white.

Epstein Island might be a little small for 80M people but they'll be ok under dictatorship...

If you want to limit the number of times you need a warrant and be right with the Constitution, you need to limit the number of people you're detaining and deporting. Stick to actual criminals rather than abuelita and protestors and you should be fine.

Gen X is mildly tickled that we were mentioned. We realize we have $800B reasons to help but our fucks were long ago depleted through neglect. When we're done with the armor of contempt we might let you all borrow it.

The anti-fascist that you're thinking about don't get elected, except in rare circumstances. And I'm speaking as one who genuinely believed that Sanders populism would have trounced Trump had the corporatist Dems not bent him over and gone to town.

*Any* fucking Democrat who we know will not try to further destroy democratic institutions or the world order. And let's be serious. If they have a reasonable chance of winning it's like choosing between a shit sandwich and a ham sandwich with pickles when you don't like pickles. Primary or General.

How one can still to this day not eventually understand how "it's all connected" is a mystery. It's essential to have people that act as the sharpest sword edge but it's more important that the hilts be wielded by people who know the best time not only to strike but when to sheath and when to talk.

This one carries less weight as evidence against GenAI's ability to summarize. Specifically, it's from 1.5 years ago and it was conducted on a single model from Meta by one of its competitors in certain spaces, Amazon. The circumstances don't make it completely invalid but it's more questionable.

ChatGPT - BBC AI Representation Study Shared via ChatGPT

I'm willing to admit my blind spots. Most of my experience has been summarization of self-contained information across, e.g., research and whitepapers, medical studies, blog posts, software docs, etc. The BBC study raises serious concerns, especially with RAG. Summary - chatgpt.com/share/6992c7...

Well OK, engage with me. I actually consider myself an "AI" skeptic (I hate the term) in the sense that I've studied the technologies enough to believe that we are nowhere near general AI. I don't think it will be achieved in the next 20 years and certainly not via LLMs.

Have you had the opportunity to upload many novellas and investigate the results of summarization? Have a look at Leon Chlon's Pythea/Strawberry project wrt hallucinations. You might be surprised.

I don't claim to know the quality of its capabilities in other domains but I do know that if one is already a domain expert, they can be quite helpful.

What work do I claim to not respect exactly? I've used it to good effect in my own career so far: as a personalized tutor for newer technologies as well as foreign languages, a sounding board during root cause analysis, building scaffolding for statistical performance analysis, etc.

Correction: as a malignant narcissist, Trump pretends to be whatever is necessary to win accolades and enrich himself. He has no core beliefs except whatever temporarily serves hi interests.

My 13 yo son would like to give it a shot. He says he has the motivation and "assets" to do it. Could you DM me some more info?

A decade longer than you, working during the days of symbolic AI and expert systems. But that's immaterial. How much precision does one need in a *summary*?

Someone who has followed technological trends in depth for a substantial period of time.

Personal finance in particular should have been made a part of elementary school curriculum decades ago.

Long time computer scientist and technologist here. LLMs are exceedingly good at finding patterns and summarizing information. Like any technology, they can be used to amplify both human virtues and skills as well as vices. Important to understand that they're not a *neutral* technology though.

The original 6 7!

One of my favorite bands of all time.

Te amo mucho!

Congress.gov | Library of Congress U.S. Congress legislation, Congressional Record debates, Members of Congress, legislative process educational resources presented by the Library of Congress

Answer this one question: did you directly feed the 3M+ redacted Epstein files to Grok (over 3 million pages released by the DOJ in 2026, plus prior batches from 2019-2025)? I see a reference to congress.gov but I can guarantee that it didn't read and parse the files with RAG.

"I've learned more about how to use AI incorrectly..." There, I fixed it for you. Something that sounds plausible coming out of an LLM does not mean that its logic is sound because it's not actually "reasoning" in any *human* sense of the word.

First, I've forgotten more about AI than you've learned. Second, Grok presents a theory in the article, without reference to prior decisions or legal analysis. Third, there's plenty of circumstantial evidence of criminal wrongdoing in the files.

Incorrect conclusion. I admitted nothing other than what I wrote. But keep engaging in logical fallacies and treating output from a probabilistic token generator as ground truth. And kudos on your new prompt engineering gig! Clearly it's working to save the few rational neurons you have remaining.

Ummm, we didn't need any evidence from the files. He's already a convicted criminal.

YouTube video by Sully_Dawg January 5, 2026

All that money has gone into his pockets.

youtube.com/shorts/j4p_S...

Screw it then, I and everyone I know will be "traveling" until one day after elections!

I think Talarico should start running now and if he loses to Newsom, he should be on a VP ticket. Crockett should run too. Hell, how about Michelle?