Notices by Dave Troy :toad: (davetroy@toad.social), page 4

📃 NEW: "Reality in Taiwan Differs from Perception in DC," my latest for Washington Spectator. While US media hints at an imminent war over Taiwan, the facts on the ground tell a different story altogether. Politics and economics, not war, are the focus in Taipei.
https://washingtonspectator.org/reality-in-taiwan-differs-from-perception-in-dc/

The fact is that it takes several thousand human hours to get up to speed on our current geopolitical situation enough to understand what's happening.

So the end result is that nearly no one is able to do that, and those who can are separated by an almost insurmountable gap from the vast majority.

This simple fact provides a significant and nearly impenetrable defensive moat for bad actors.

Some observations about LLM's:

• Math: Can't really do math, doesn't have any model for it
• Research: Makes stuff up, can't be trusted, is extremely confident
• Coding: Unless it hits the correct example somewhere else, will usually get it wrong
• Data Analysis: fakes it until it makes it, but breaks down at any kind of scale
• Writing: as likely as not to plagiarize in ways that are hard to predict

If this was an employee or intern, I'd fire it. ;)

1/Please note this slide from a Mike Flynn event this past weekend, calling for "US dollar crash, gold & silver multiply" as part of unlocking the "third seal of Revelation."

Flynn and Bannon have been nurturing this strategy for at least the last two years, and they see debt default as the path to fulfill this prophecy.

1/Suppose for a moment Karl Marx had not been born.

It is very likely that the 20th century would not have played out as it did. I don't mean to pick on Marx specifically; but individuals induce ideologies, which induce social ties that produce history.

There is nothing particular about Marx's articulation of his ideas that were inevitable; they reflected his observations in the context of his time.

Suppose instead there had been a Dutch writer named Van Halen who published a theory...

2/of capital and social organization that had taken hold. There would then be a bunch of "Van Halenists" in battle with some other group.

Maybe the Austrian Artist got into art school. Or Ayn Rand had stayed Alisa Rosenbaum and never made it out of Russia. The 20th century as we knew it would have played out very differently.

Some might argue the same forces would have come to blows, but it still would have played out differently.

All this is to suggest that ideas and timing matter immensely.

3/So as we are dealing with our current battles, it is important to recognize just how provisional and tentative they really are: they are frameworks that are totally accidental and of our moment.

We reify them in our minds until they take on the strength of steel beams. But they are nothing more than gossamer threads that stitch reality together in a framework we can relate to. They are a set of stories we can reference together, and within them battle for advantage.

4/But less explored is the value of new stories, new ideas, and new accidents — all of which are available to us at any moment.

We can invoke them anytime we like, and there is an infinite supply. Why must we do battle within the stories that have come before, instead of inventing and telling new stories?

Steve Bannon famously said "flood the zone with shit." Why should we not "flood the zone with heroes?"

There is no shortage of new North Stars.

My assessment from November 20, 2021 still stands.

"Bottom line: Peter Thiel and Elon Musk have allied with Putin and oil/gas interests to try to undermine the dollar as the world’s reserve currency, in a fascist bid to destroy the US and EU administrative states.

Pick a side."

#ChatGPT is impressive but uninspiring. Yes, it possesses a facility for filling a page with text its creators stole from human authors. But it is, by definition, unoriginal and lacking in thought or creativity; a parlor trick, with dead eyes.

Anyone who has experienced “Eliza” or Markov chain text generation (both have been around for ~50 years) sees that this is but a very very good implementation. But it is rooted in theft of human creative effort; an avatar of today’s Silicon Valley.

For those interested, this is a pretty good Mastodon server setup:

1) hardware box w/32 cores, 32gb, 1TB SSD, 1Gb eth, Ubuntu 20.04,

2) 8x Puma, 1xPgsql (pgtuned), 1xRedis (500mb, no save),

3) Sidekiq in 4 procs, 25 threads/db-conn, with pull, ingress, and default each with a dedicated process.

I think this is going to scale quite well; can add more sidekiq procs. And then will break out to multiple boxes. But lots of dedicated CPU is the key bottleneck. This slaughters the queue.

@ken_sugar Dorsey says “nope.”

https://twitter.com/reuters/status/1593155854987042816?s=46&t=wYXjWt38uxRGv_A7XFxkVA

A few recommendations for life on Mastodon:

1) Follow anyone you think looks potentially interesting; you can always unfollow later, and they may lead you to new people via boosted posts.

2) Boost posts you think are worthy, so others can discover new content.

3) Don't obsess on replicating your Twitter follows on Mastodon; let it be its own experience, and grow it organically. Obvs follow anyone you miss from Twitter, but this isn't a 1:1 replacement; have fun, follow your instincts.

Another design consideration re: Mastodon is that it works well for ephemeral asynchronous communications, but for many reasons should not be counted on as an archival resource. Media attachments are periodically purged and may not be available after a week, or a month, etc. While some servers may try to preserve content forever, this may be costly and unsustainable. Creators, researchers should treat this as an ephemeral resource and make provisions for self-archiving anything important.