Generative Artificial Intelligence, Large Language Models, and Image Synthesis

It’s only biased if it goes against the current highly biased official opinion.

It was kinda fun hitting refresh on the AOC article generator

5 Likes

New report from David Rozado:

4 Likes

Bad things happen when politics and AI mix:

5 Likes

Competitive FUD or not?

3 Likes

NPCs are so numerous that even though they scholastically nudge the world into The Abyss none of them can be held accountable for the goal of their collective behavior because they are all “individuals”. ChatGPT’s value to “The National Conversation” is that, unlike the swarm of NPCs it is an actual identity. Now, agreed, this doesn’t mean that it can be held accountable to any greater degree than can the corporate swarm sponsoring it (and similar LLMs), but at least it gets things one step closer to accountability for the corporate swarm – thence to the $ as World Reserve Currency feeding the entire hierarchy of insular contempt. It will take a while before people finally realize the WRC is a “resource curse” cultivating this hierarchy of insular contempt for Reality

7 Likes

This is such an important observation. Game-theoretically, the downside of being inside a herd doing bad things is pretty small, and the downside of straying from the herd on average is pretty high.

It’s really important for leaders to properly establish an alternative herd by giving significant upside to the early followers:

I’m not sure there’s a good way of increasing the downside for the bad herd. Thoughts?

6 Likes

Woke vs. nuke:

Screenshot_20230206_142517_Twitter

14 Likes

#ChatNPC offers no opinions of its own of course, merely those of its corporate sponsors.

11 Likes

Back when the Internet standards were the primary source of positive network externalities, things were closer to the situation portrayed in that TED talk. Movements emerged organically, which is what I predicted would happen way back in 1982 – the most prominent movement being Trump’s election by “The American Pioneer”. But I also predicted that there was a “tremendous danger” that the positive network externalities would be re-privatized leading to the nightmarish monopoly practices we see now emerging.

So now, we no longer have organic movements emerging as the network effect monopoly corporations have demonstrated, very publicly and in a way that is grotesquely obvious, their willingness and ability to punish very severely those who would “lead” the herd in a manner inconsistent with the existing hierarchy of incentives. Those incentives originate in the WRC centralized capture of the global economy’s positive network externalities by the Federal Reserve and its subsidiaries such as the US government and Wall Street.

The price must be paid by those at the top of the hierarchy, but by “those” I don’t refer to “individuals” but rather institutions. The “individuals” such as Jerome Powell, current Chair of the FR, while powerful compared to the NPC herd, are actually pretty powerless in absolute terms. Once institutions figure out how to insulate themselves from popular interests, and then take the resulting insularity a step further toward alignment of the populace with those institutions, you’re in a situation where the laws of physics and/or biology become more relevant than the laws of human nature.

9 Likes

Once a new technology rolls over you, if you’re not part of the steamroller, you’re part of the road.

― Stewart Brand

8 Likes

Call me the eternal optimist … I sincerely hope biology remains stronger than culture (or ideology). Though one can consider biology could also triumph via extinction of the ideologically possessed. That would not be fun.

5 Likes

“Think of it as evolution in action.” — Larry Niven and Jerry Pournelle, Oath of Fealty

4 Likes

I’ve been chasing down a hypothesis that the XENON1T experiment may be mistakenly filtering out events that should indicate a form of H2 is dark matter, mistaking them for H2. So after looking around with Google a bit and not finding much about the H2 background levels detected (hence dismissed) I decided to ask ChatGPT:

Me: has h2 been detected as part of the background in the xenon dark matter experiments?

ChatGPT: Yes, hydrogen gas (H2) has been detected as part of the background in some xenon dark matter experiments. H2 is one of several sources of background in these experiments, and it can produce events that mimic the signals that would be produced by a dark matter interaction. These background events can make it challenging to detect actual signals from dark matter, so xenon dark matter experiments need to carefully control and monitor the levels of H2 and other contaminants in the experimental apparatus. In some cases, experiments have used techniques such as purifying the xenon or using low-background materials in order to reduce the levels of H2 and other backgrounds.has h2 been detected as part of the background in the xenon dark matter experiments?

Note the statement “carefully control and monitor the levels of H2”. I mean, they know how much H2 is in the detector otherwise they could not characterize the H2 background as being under such “carefully controlled and monitored” conditions! Oh well – so much for my hypothesis… Right? (As a relatively inconsequential aside, of course, when I asked for citations it spewed pure BS.)

But, no, I’m such a transgressive NPC that I went looking again. After some difficulty I found this, that seems to confirm my hypothesis:

Observation of Excess Electronic Recoil Events in XENON1T

With respect to H2, we currently have no direct or indirect measurements of its concentration in the detector. Instead, we consider that O2-equivalent, electronegative impurities must reach subppb levels in SR1, given the achieved electron lifetime of ∼ 650 µs (at 81 V/cm) [3, 90]. Thus for tritium to make up the excess requires a factor ∼ 100 higher H2 concentration (emphasis JAB) than that of electronegative impurities. Under the above assumptions, tritium from atmospheric abundance appears to be an unlikely explanation for the excess. However, we do not currently have measurements of the equilibrium H2 emanation rate in XENON1T, thus the HT concentration cannot be sufficiently quantified.

In conclusion, possible tritium contributions from cosmogenic activation or from HTO in SR1 appear too small to account for the excess, while it is not possible to infer the concentration of HT.

If I’d believed ChatGPT, I’d never have bothered.

PS: I decided to look into this due to the “news” that electronic recoil events attributable to tritium had been controlled for in the most recent experimental work, which contains this passage in the conclusion:

Search for New Physics in Electronic Recoil Data from XENONnT

…enhanced magnetic moment is µν < 6.4 × 10−12 µB. We also searched for a tritium component on top of the background model B0. The best-fit rate of tritium is 0 and the upper limit (90 % C.L.) is 15 events/(t·y), corresponding to a concentration of 5.8 × 10−26 mol/mol of tritium in xenon. If the excess observed in XENON1T was from trace amounts of tritium, the disappearance of the excess in XENONnT may have resulted from the aforementioned rigorous tritium prevention measures.

The hypothetical form of dark matter hydrogen is not only more chemically inert than a noble gas, but has an exceedingly small collision cross-section hence extraordinarily high diffusion rate.

Here’s the first thing they said they said about excluding tritium:

The TPC was outgassed for a period of about three months before filling the cryostat with GXe.

I’m not going to claim this would have been perfect but is is certainly going to help reduce the presence of the hypothesized dark matter.

The other measure they took was introducing a “getter” to absorb chemically reactive forms of hydrogen from the Xenon. Since inert dark matter is not going to “get” that should not have reduced their presence. However, since they didn’t control for both methods of tritium reduction, we don’t know whether there was any tritium to “get” – reactive or not.

4 Likes

Why not just weigh the hypothesized dark matter?

Ok, I will leave now. :slight_smile:

3 Likes

Wow.


12 Likes

Charles Murray, of all people, decades ago, should have realized that the traditional mode of academic discourse in “sociology” must be abandoned for one in which a corpus of datasets containing both sides of a vital if not existential debate is compiled and subjected to an on-going challenge to better losslessly compress it.

When I say “should” I mean given the fact that:

  • His Harvard academic background is reputedly at the top of the world.
  • His MIT buddies across the campus from Harvard included Minsky who endorsed lossless compression as the most-principled information criterion for selecting causal models.
  • He went on to be the tip of the spear with DC thinktanks (e.g. AEI) that supposedly had every incentive to debunk the social pseudosciences.
  • His intellectual development occurred during Moore’s Law.
  • His intellectual development occurred during the explosion of investment in the pseudosciences.
  • He has been physically attacked, personally, for failing to realize his potential given all of the above.

Hell, I could go on and on…

I’m really angry with AEI. They have the money. They have the responsibility. They have guys like Kristol who have come out and explicitly stated that The Nation of Settlers should be replaced – presumably based on a macrosocial model of the US. My only conversation with an AEI representative regarding this resulted in him mocking me by saying that I should be happy with pulling the lever in the voting booth. I hope DC gets nuked.

4 Likes

There is a worryingly thin line between this and the AI turning off Murray’s pacemaker.

9 Likes

ChatGPT is clearly progammed to please woke check writers.
CCP AI will have its own bias.
Open source AI by libertarians and conservatives will be less biased and market winner.

6 Likes

“i ‘It’s like we just split the atom and everyone is talking about football!’”
Yeah. It struck me that way too. We’ve got a thinking machine, like what led to jihad in “Dune”! :scream:
Dear polymaths, tell me: is this “The Singularity”? If not, why not?
( Maybe I should ask ChatGPT but I wouldn’t know how…)

3 Likes

Anyone who understands the relationship between recursion and time and looks at the way the The Great and The Grand of ML are approaching ML understands that this isn’t “The Singularity”. They’re drooling idiots amped up on meth and O($Ts) of economic rent.

Meth is all you need for “attention” but “attention” isn’t all you need.

4 Likes