Apple has not sold directly in Russia since 2022. There are parallel imports but no direct sales.

You’ve only described one half of the coin (and not very accurately in my view). The other side is the flip side, under which there are huge quantities of CSAM, suicide being promoted, scams and hustles of every kind tricking or coercing the vulnerable into giving up their savings, and of course endless torrents of the vilest abuse directed towards lots and lots of people of various types. The problems of the internet absolutely do include people saying things that would not be legal if said in any other medium, whether on TV or in a pub.

Your first post sure made it sound like you had clear insight into what these folks were and were not predicting. This second post makes it sound like you don’t know. Until you know, how can you tell whether or not what they’re providing is useful or not?

I feel like we are still living in the golden age of AI. Like USENET in 1993 or Netflix in 2020.
AI isn't perfect, but is incredibly helpful in many areas such a tech and education, though not all, and at least tries to be helpful.

Future problems:
- The trillion-dollar market caps of the infrastructure companies (like NVIDIA) and the multi-billion dollar private valuations of the pure-play AI models (like OpenAI and Anthropic) are only justifiable if the ultimate revenue stream is measured in trillions per annum. There will be pressure to generate revenue by any means possible.
- Content creators, SEO firms, and malicious actors know that LLMs are trained on public internet data. They will flood the web with AI-generated, manipulative, or false content specifically designed to influence future models.
- If Generative AI content dominates the internet, the models will start training on their own output, creating a feedback loop of consensus bias without grounding in real-world facts

The ultimate risk is that the fundamental utility of the AI assistant—its ability to be an objective, reliable source of information and help—is compromised by a shift in its primary allegiance from the user to the corporate bottom line.
(this post assisted by Gemini 2.5 Flash)

...In all those millions of driving miles? That's a far better record than human drivers in any single small town.

The entire positive side to bureaucracies and committees and governments is that they have enough people in them to do multiple things at once.

Usually when someone says something like what you said and I quoted above here, they are trying to argue that human drivers shouldn't exist. Maybe this is true, for some particular set of truths, but there's always a number of ways you can look at a situation. For example, I would argue that no one and no computer should be driving in the bulk of situations we are currently driving in, because cars are a terrible mode of transportation in the cities where most people live.

That's true, but that doesn't make what you said above correct. Left and right are very, very different things, and it very much does matter which you get.

Isn't it a fake race with fake metrics? We have companies sinking many billions of dollars into some vague development of AI, but without concrete goals or definitive plans to turn that investment into revenue.

I guess if someone wants to declare themselves the winner, they'll simply have to be the one that burned the most capital on this boondoggle.

First Street very likely doesn't have some magic model that can predict the future better than anyone else.

When you get a mortgage you have to pay for a flood survey. Even my house 700' above the village where the bank is.

Your flood risk is absolutely predicted by the flood history of your location. The bank writing the mortgage has the skin in the game which is why they make the buyer pay for the flood survey.

It sounds like First Street might be liable for damages based on pseudoscience if these Realtors bring a case. It would be interesting to see them present solid evidence that they prospectively beat the existing flood models and survive a cross-examination.

If they've published a peer-reviewed paper then I missed it.

Hostile design is often a sign of libido dominandi, not just laziness.

SmartTube has probably hundreds of settings you can tweak to improve usability and accessibility. The developer clearly has a user-first philosophy.

You have Revanced for Android but Apple doesn't allow such things.

It has some excellent accessibility improvements over stock.

Cheaper to just pay the bribes.

In America it's known as K-street. Or "donating" to an Inauguration Gala. Or hosting a high court judge in a European palace for a couple of weeks. Or giving decision makers absurd private sector salaries when they 'retire'. Or giving the Governor's wife a $200K no-show job. Pick your branch, there's a way.

In India the system is less formal.

Look, if someone makes a mistake abd owns it. We learn from it and move on. If your business is destroyed by a single person's slip up, then perhaps you are not that serious of a business.

The absolutely toxic corporate culture that firing someone is a reasonable first step is why millions of office workers are paralyzed with fear over losing their job. You are not trash to throw away, you're a person that society has invested thousands of dollars and hours into childhood and education.

In reality, people should be let go if there is a pattern of behavior that other measures have failed to correct. Like if they aren't completing tasks, behaving in an unprofessional way, making costly mistakes frequently, etc.

We're still in the dark ages when it comes to social development. Humanity has a long way to go before we can competently wield the power that we now possess.