on some black I-80 mile

1. The court holds Google responsible for statements made by its AI, considering them Google's statements (search engines have limited liability for results in their engine as they're the words of other sites/companies/people), meaning when their AI lies/hallucinates they're liable for the defamation/harm resulting from those statements.

2. Google's defense that customers are generally aware of the lack of reliability and are responsible for fact checking was dismissed. As the court pointed out, that would "significantly diminish" AI Search's stated purpose and it can't be distinguished from Google's business practices/statements as a search tool.

3. Studies have found about 91% of Google's everyday AI responses are accurate, leaving millions of searches per HOUR with potential liability for falsehoods. 56% of correct responses weren't supported by the sources the AI listed. Both of which mean Google is now liable for a LOT more AI "errors."

4. Google was held liable for 80% of court costs in this case and this precedent is expected to reverberate around the world. This is a massive shift from the 3rd-party search provider role Google has previously played and it comes right as they've tied ALL searches to their AI search.

I had to turn this off, but it's something that allows Windows and anyone using your device to generate text/images.

I remember trying to look up clover references, so I searched clover. Prior to the update it would have given me photography and stock images in abundance, interspersed with art, and I could further limit my field to be ONLY photography or stock images. This time around: it gave me Clover, the character from Totally Spies, specifically involved in inflation art, because that's what the algorithm decided to push at me since it wasn't massively popular and it clearly needed more attention. I couldn't narrow my search, because the references I wanted were old and thus had no keywords. Mass-uploaded photos by amateurs and professionals alike often just had numbers for a title. There was no way out. It was just me against the Clover inflation art.

This is a form of backdoor bigotry from both normative facts (there are more straight than gay relationships) and well represented bigoted beliefs (men are superior to women).

Carolina Milanesi, an independent technology analyst, said Google is trying to make its cash cow business — search — richer and more personalized, and it will make shopping easier. But there is a risk that users may have fewer choices about what to click. "Right now it's: I ask a question, I get a bunch of answers and I feel that I'm in control as to which answer I take, or if I'm looking for something, which product I'm going to end up buying. That is going to be less so going forward," she said. Milanesi envisions AI-enabled search and agents proposing products to consumers — perhaps even those they have requested — but with less clarity or choice around where it's coming from. "If you're going to say: 'I want a pair of Jordans, go find them,' you're not necessarily sure what steps have been taken and whether the AI has used a source or a store that was paid for and therefore came up in the search results," she said, "or if AI actually went and did their due diligence and picked the best for me as a customer."

And here's one from Time magazine (May 20, 2026):

While Google already has “AI Mode,” the company will now power the whole search bar through its new Gemini 3.5 Flash model.  Instead of the classic list of blue links, Google Search will now also generate a custom page with an AI-generated summary of what you’re searching about, which will then trigger a conversation with AI Mode on the main page, allowing users to ask follow-up questions—similar to the kind of layout you would see when opening ChatGPT.

And a little more from Time's article on how this may affect the websites that we are trying to search for:

Let's be precise about what @queercodedangel got wrong.

What the IRD list actually was:

A private notebook, handed to one person at the IRD, listing individuals Orwell thought were unsuitable to write anti-stalinist propaganda - not people he wanted arrested, surveilled, fired, or harmed - and none of them were.

Orwell believed these specific people were Stalinist sympathizers who would undermine the work of countering Stalinism. You can legitimately think that was a bad call - it was certainly a morally complicated one for Orwell. It was not, however, the act of an imperialist snitch. It was the act of someone who hated authoritarian Stalinism enough to do something uncomfortable about it, while continuing to publicly criticize the British state.

The tension between those two things evaporates the moment you understand Orwell's position: Stalinism wasn't socialism. It was fascism with better branding, and it was destroying the socialist project from within.

History proved Orwell right.

Orwell:

Got shot through the throat while fighting fascists in Spain

Spent decades producing some of the most devastating critiques of British imperialism ever written, including Burmese Days, and Shooting an Elephant.

Called himself a Democratic Socialist his entire adult life and meant it.

Wrote that "every line of serious work I have written since 1936 has been written, directly or indirectly, against totalitarianism and for democratic socialism."

@queercodedangel is taking one late-life episode, stripping it of all context, and canceling one of the sharpest anti-totalitarian, anti-imperialist pens of the twentieth century - because he wouldn't excuse Stalin's crimes. That's apparently the price of admission to the 2026 left - being ignorant enough to deny that Stalin was a monster.

Calling that "snitching" tells you nothing about Orwell. It tells you that for some people, anti-Stalinism is a greater sin than imperialism ever was - and you can only take that position if you know nothing about what Stalinism was.

That sucking vacuum of ignorance where actual knowledge should be doesn't inspire a single second of hesitation before they confidently preach provable falsehoods. The less @queercodedangel knows about Orwell, the IRD list, or Stalin's crimes, the more confident the verdict. This is what happens when a generation decides memes and reels are sufficient substitutes for actually reading anything.

Orwell spent his life fighting this kind of epistemic laziness - the willingness to flatten complex realities into politically useful, emotionally resonant, self-affirming fictions.

This is why Gen Z desperately needs to read the Orwell and Huxley they've been told to avoid.

Little website tracking the expenses and revenue of the big AI companies, and literally the only company making money is NVIDIA because everyone pays them for data centre hardware.

NVIDIA ^

Let's illustrate : Amazone has invested 313 billion dollars since 2022. It has recouped 22 of these billion dollars since then. That means a total of 8.8 billion dollars per year (We're in June 2026, so I'm assuming 5 semesters, so total earnings divided by five then times two for the yearly amount). That means that their RoI is, in percentage, calculated as follow : 8.8 / 313 * 100 = 2.8%

That's not a very good ROI. In years term, that means it would take them almost 38 years to fully pay off their investment ... if it doesn't increase in the meantime.

For Alphabet, the ROI is at 3.5%, which would require almost 28.6 years to pay off. For Microsoft, the ROI is at 4.6%, for 21 years and 9 months to recover. And Meta manages a ravishing 0.5% ROI, for a total of two entire glorious centuries to be repaid.

Now, low percentage ROI projects aren't necessarily a bad idea, you'll find plenty of them in industry, like for instance for acquiring new machines or vehicles, or creating a new production center. The main, biggest issue here is ... those investments are made on AI development and data centers, which are respectively : - A bundle of lawsuites tech gadget fad that only survives due to a constant train of hypemen being run through the news 24/7 (either gushing about the Glorious AI Future or screaming that the End is nigh, y'all) - Buildings that turn costly, medium-lifespan equipment like CPUs into short-lived consumables.

If you're investing in a new production unit to make cookies, 4.6% ROI can actually be pretty damned good. With AI, it's ... not as good. For multiple reasons. For one, like I mentionned earlier, they're on an investment-as-a-service kind of plan, where they have to pay more every month to keep investing into AI, which means that the ROI is actually going to look worse and worse as time goes on, unless they somehow manage to convince enough users to put more money into buying AI services than they put into keeping those services alive.

For two, that's while AI is booming, is having much success. That's ... as good as it gets. It's actually a feature in a lot of industrial projects that, well, they bring in comparatively less towards the project's end.

The only way AI looks like a good investment is if you include stock market valuation, which is terminally irresponsible and would drive any serious accountant to fits of rage and/or fury, because that's not real money, that's not an actual cashflow, that's imaginary wallstreet casino chips, you don't pay employees and suppliers with that.

Which leads to NVIDIA, who certainly seems to be making off like a bandit with this .... except they are de facto a self-made hostage. If those AI companies tank, NVIDIA does too. Big time. Also, I'd be willing to bet those 253 billions include the 100 billions NVIDIA is reinvesting into OpenAI that OpenAI then used to pre-buy chips from NVIDIA.

So, conclusion : ROI is a measure of how profitable a project will be yearly, and can be used to estimate how many years need to go by before it has paid for itself, AKA how long until it actually turns a profit. Companies and banks typically prefer projects that are on the shorter side and have good ROI to limit risks. Long lasting projects will typically require a material, physical element that can be seized and resold to be supported. AI ... AI's horizon for repayment is past 20 years in the best of cases, which is more typical of heavy and complex infrastructure projects, the kind that will be in use for half a century at minimum before needing notable refurbishment.

No! The problem with feckless liars is that they don't give a shit, period.

He's not like Manchin or Sinema or even Fetterman. He's our answer to George Santos. He could start a cult tomorrow. He could shoot someone. He could quit. He could disappear.

People who lie constantly cannot be trusted in any capacity. They can't be given responsibility. They are at baseline not good people. They are not accountable to reality. They don't respect anyone or anything.

They believe that they can evade consequences for bad behavior by telling more lies. They become more volatile when they are under stress. They double down when the stakes rise. When it finally becomes clear that lying won't work, they melt down.

They even damage trustworthiness in general! They create a haze of distortion, because once they're part of the group people will rationalize their lies. "Maybe he didn't know he got a Nazi tattoo!" "Maybe he has worked on his marriage!"