perestroika

"More fundamentally, AI models may not understand ‘stakes’ as humans perceive them."

In my repeated attempts to solicit the advise of various language models for some situations which a programmer might face (e.g. being unable to read all the world's literature of a subject), I have come to conclude that they cannot understand "truth" as humans perceive it. Today's language models don't fail apologizing, stepping back or admitting inability - they fail confidently bluffing.

Possibilities:

• their training material does not include enough cases of humans apologizing about being unable to solve a problem
• a bias was introduced to get them to ignore such cases, since admitting such material resulted in too frequent refusal or self-doubt

Basically, today's models seem to be low on self-criticism and seem to have a bias towards believing in their own omniscience.

Finally, a few words about the sensibility of letting language models play this sort of a war game. It's silly. They aren't built for that task, and if someone would build an AI for controlling strategic escalation, they would train this AI on rather different information than a chat bot.

All I can say is that they have to find a solution, and their time is limited.

I think it's a safe bet to say that Japanese sociologists are already studying the problem for decades.

Choosing some solutions they might propose is becoming increasingly urgent. It will most likely require adopting a different view on topics like employment and career. If a career is valued highly and having kids blocks it effectively and incurs enough expenses to set a person back in life, then people are discouraged from raising kids.

It does seem that Taliban has performed various attacks on Pakistani troops in recent past. If negotiations failed, then a counterattack would be expected.

However, a war between the countries seems unlikely. Afghanistan does not have an economy capable of sustaining war, and Taliban does not have a well-equipped army. Meanwhile, Pakistan is heavily equipped.

Also, Taliban is absolutely certainly not representing the majority of the Afghan people (for example, one would struggle to find a woman supporting the harsh removal of women's rights). The majority of the population most likely don't want a war, and Pakistan most likely doesn't want to conquer a foreign land.

I think this will end with Taliban's command centers and weapons stores being bombed, after which negotiations start succeeding.

There have been reports of Taliban's political leader Hibatullah Akhundzada dying in the bombing raids. I have not seen enough confirmation to say it has happened. If this is true, it would be a bit more complicated, they would have to choose a new leader.

One of Uganda's main sources of foreign revenue would be coffee.

Companies might also want to check where their cobalt, gold, tungsten and copper comes from, since some part could be coming from Uganda. Remembering the name "Kilembe mines" and avoiding any dealing with them would put a dent in Uganda's budget.

Personal opinion: boycott Uganda as long as their president keeps opressing minorities.

It's a short step from opressing sexual minorities to opressing religious minorities. From there to ethnic minorities: just a small step. From there to repressing policical opposition: another small step. From there to something incredibly harmful to the whole people: just a small step.

I'm surprised the Druzhba pipeline still occasionally functions, despite Russia fighting a war against Ukraine for what is now 4 years.

Orban should not expect Druzhba to keep functioning forever. It will not be blown up in Hungary. It will be blown up, repaired, blown up, repaired, and finally blown up and no longer repaired - in Russia and most likely, also in Ukraine. There are no guarantees for its survival. All the alarms have been ringing for 4 years. If Orban wants oil, he should buy it from elsewhere than Russia. He has been offered oil through Croatia, it's just not Russian oil. He should accept the offer and forget about Druzhba.

As for Orban himself, if I was a Hungarian, I would get rid of him. He's likely setting records of corruption and media manipulation right now, trying to manipulate his way to re-election for the umpteenth time.

bankrolled by the US to further their nefarious interests in the region.

A long time ago. In response to the USSR couping and occupying Afghanistan. Which was, of course, also not justified.

If Afghanistan had been left alone, things would very likely be better in several countries...

The honorable tradition of fragging officers and rebelling as cohesive units certainly needs to develop in Russian society. Its relative absence tells me that anarchist-minded people are abroad or hiding - none of them are in the army.

The army, as things look, is composed of 3 main demographic groups:

• a minority of professional soldiers loyal to Putin's regime (they attack with actual equipment after meat waves have opened a gap in Ukrainian defenses, they also run the high-tech weapons)
• simple and stupid blokes, often a bit enthusiastic about Putin's regime, bribed with money or tricked to enlist
• detainees and convicts promised freedom if they join the war, foreigners tricked to enlist, Ukrainians seized from occupied areas

The latter two groups get used as cannon fodder. About 30 000 per month.

All the groups are slowly dwindling in numbers. Loyal goons are increasibly dead, a military career looks less glorious if you hear of everyone who died. Society is running out of the really gullible, and even the really poor often prefer life in poverty to a high probability of death without ever seeing anything different. As for convicts, word has sufficiently spread in prisons that coming back is unlikely, so they're volunteering less.

On this background, the possibility of a mobilization looms on horizon. But Putin fears giving weapons or ammo to randomly chosen people. They might rebel or his house of propaganda and cards might collapse.

I recommend reading the article. It's depressing, but clear.

These guys, mostly, were killed for refusing to go on suicide attacks. Some for retreating. Some for refusing other orders. Some, possibly, for simply existing in the wrong place. The unlucky ones, for failed desertion (I hope that lucky ones outnumbered them and deserted successfully).

Women and children, except for those who got caught on the wrong side of the front, fortunately, aren't the typical casualties in a war with a static front. They get evacuated before the front crawls near. But they do get killed when glide bombs, missiles and drones are targeted at random buildings.

Please don't lump Afghanistan and Iraq together.

Afghanistan had no oil. The US invaded because it was attacked by a group based in Afghanistan. The end was a mess, but the invasion was justified. This has been the only time NATO article 5 has been activated, by the way.

Iraq has oil. The US has invaded twice. Once because the UN authorized restoring the independence of Kuwait, after Iraq had invaded Kuwait. The second time because emperor Bush II lied to everyone about weapons of mass destruction. The end was a mess, of course.

Does this have any predictive value for Iran?

I can only tell this:

• I see no preparations for a land war
• regimes may sometimes fall after suffering a military defeat
• only one regime has fallen recently because of an air intervention (Gaddafi in Libya, but the Libyan civil war had already started at this time)

As a result, I am not convinced of any outcome, I do not support and do not oppose, I observe and try to understand.

I will try to explain. Protesting allows to determine how many are generally "on board" with an idea. If you protest and you see that you have a supermajority (e.g. 90%) the point is moot, then the armed services are likely also on your side (e.g. USSR in 1991).

The next step, changing the government, usually requires a bit of force. It can take two typical forms:

a) people overcome cops while the army does nothing

b) units of the army rebel and overcome cops

c) units of the army fight each other in a civil war

Now, as you may guess, option C is pretty bad - weaponry made for international defense gets unleashed internally with no clear lines of conflict. But option A is pretty tame and B can be reasonably quick and bloodless.

It all starts with a show of hands, and an evaluation of how many are OK with a plan - inside a group and between groups. If too few show up, one knows the time is wrong. If a modest amount show up, one knows to be careful. And so forth.

Some thoughts:

1. Social media encourages a herd mentality.
2. Social media can distract you from learning and focusing.
3. Social media can feed you disinformation or subject you to bullying.
4. This applies doubly for algorithmically steered social media (where information gets pushed at you without your request).
5. Algorithmic social media is good at getting engagement (conflict: just another form of engagement).

This is being done to earn someone profit, not to inform people but maybe entertain them a bit - but foremost, because someone buys advertisements on these platforms.

Courses of action:

• deal with the users? (carrot or stick?)
• deal with the companies?

Viability of different courses:

• Users are diverse and numerous. Dealing with them is a lot of work.

• The actions of malicious users may be an offense in some place, but not in another, and being rude is not an offense. Dealing with malicious users individually is feasible only in an environment where the general public is interested in peace of mind and orderly discussion. I have seen and still participate in such places. They are mostly non-profit forums wich a few clear administrators. Lemmy is an experiment in a similar direction. An algorithmically steered social media site run by a for-profit company... no, it does not fit this profile.

• Education of non-malicious users - how to choose a good environment and defend oneself and others in this environment - may be more effective. Users should be informed about what benefits them. They should know that "environment A has effective moderation, while environment B is a troll cave". They should know that "environemnt A sends you information that you asked, while B pelts you with rage bait". Who should educate people about the environments they can choose? Obviously, schools.

• Very large companies have to comply with EU DSA rules. They must show how they have effective moderation, prevent hateful content from propagating, are not harming minors, etc. I would prefer if all large companies were pressed towards effective moderation. It hurts their profit margin, but they must accept this is the price to pay for operating in a civilized society.

• A particularly blunt instrument, recently touted in several countries, is outright banning of all underage people from social media. This is highly controversial. To the companies it sends a message: "if you cannot create a safe environment, we will take your future customers away, full stop". To others, it causes great inconvenience due to age verification, which is problematic. If age verification is in place, using social media for publishing something pseudonymously or anonymously becomes near impossible. I would not like that to happen.