Copilot was uniquely awful at this, because up to until literally days before the switch to usage based billing there was no way for people to track token usage, despite repeated calls from the community.
Microsoft only added a billing “projection” feature on the admin page that was meant to download a spreadsheet (which straight up didn’t work for most people) less than a week before the new billing structure.
Doom is Turing complete FYI, so if it wasn’t for the inherent limitations on the number of gates due to small map size Doom would be able to run Doom
But can DOOM run it? - https://calabi-yau.space/blog/doom.html
From the manufacturer’s point of view all they see/hear from big tech is “these fridge ads are very effective, see this guy who bought a fridge as evidence.”
Not just influencers. EdTech, including Khan Academy, is pushing AI to make virtual versions of historical figures to engage students.
Because wealth rapidly becomes illiquid as it increases in value. The capital of the ultra wealthy is intimately tied to physical assets in the real world that makes it difficult for capital flight to occur. They are wealthy because they own the supermarkets, the hospitals, and physical infrastructure.
An example of this is when the UK froze/seized Russian billionaire Abramovich’s assets as part of sanctions, which included the Chelsea Football club. Abramovich (and the media) made a big stink about it, but was ultimately forced to sell.
The case for keeping the ultra wealthy around for their entrepeneurial innovation is inconsistent with broader economic evidence. Bell et al 2019 (QJE) show that although it is true that the young of the top 1% (again, not the ultra wealthy) are 10x more likely to become inventors, the actual causal mechanism is exposure effects, i.e. having strong peer networks.
This is something that is explicitly worsened by abolishing wealth taxes, which are well documented to increase wealth concentration and social mobility.
Who Becomes an Inventor in America? The Importance of Exposure to Innovation* | The Quarterly Journal of Economics | Oxford Academic - https://academic.oup.com/qje/article-abstract/134/2/647/5218522
I suggest you read Zucman’s common wealth tax objections, which addresses most common rebuttals.
https://gabriel-zucman.eu/files/saez-zucman-wealthtaxobjections.pdf
I applaud the high quality references, but the wealth taxes studied by Econ are often not the wealth taxes that are in popular discourse.
In particular, the general public, and the few inequality economists like Piketty, Saez, and Zucman, advocate for a billionaires wealth tax. Due to the lack of any sort of data on this, the broader economics profession generally only studies the effects of wealth taxes in the top bracket, which targets the wealthy, but not the ultra wealthy.
The NBER paper you cited is particularly egregious. The claim that a one percentage point increase in the top wealth tax rate leads to capital flight is arguably misleading because Sweden and Denmark have infamously flat taxes (contrary to popular belief). Their top wealth tax applied to singles/couples with a wealth exceeding 1500K/3000K, or about 150K/300K USD.
That is… not very wealthy, and if anything only middle/upper middle class. Not at all what a billionaires tax would target.
The broader evidence is that Sweden has exhibited a serious backsliding in its inequality measures, and the collapse of its welfare state after abolishing wealth taxes.
‘We got lazy and complacent’: Swedish pensioners explain how abolishing the wealth tax changed their country - https://theconversation.com/we-got-lazy-and-complacent-swedish-pensioners-explain-how-abolishing-the-wealth-tax-changed-their-country-272041
Ditto. I don’t see how this is different/better from existing harnesses such as Opencode, Pi, and even “commercial” open source offerings such as the CLIs for Codex, Copilot, and Gemini, especially once tricked out with plugins and extensions.
Khan Academy has been on the enshitification journey for a while now.
Sal Khan in 2023 proudly proclaimed that AI could revolutionise EdTech with highly personalised tutoring, and made Khan Academy go all in on AI with their Khanmigo AI assistant tutor. Part of this ted talk explicitly has the AI pretend to be historical figures for assistance in history.
Three years later, Sal Khan admitted what all teachers already knew: that the AI tutor failed because most students didn’t even both with it. Digging through forums reveals the similar suite of reasons: factual inaccuracies, inability to understand basic arithmetic, and opaque pricing based on “usage.”
Didn’t stop him from partnering with big tech to offer $10K bachelor degrees though.
Sal Khan: How AI could save (not destroy) education | TED Talk - https://www.ted.com/talks/sal_khan_how_ai_could_save_not_destroy_education
Why Sal Khan is rethinking how AI will change schools - Chalkbeat - https://www.chalkbeat.org/2026/04/09/sal-khan-reflects-on-ai-in-schools-and-khanmigo/
This CEO has teamed up with Google, Microsoft, and McKinsey to build an AI degree that could rival Harvard—and it will cost only $10,000 to attend | Fortune - https://fortune.com/2026/04/15/sal-khan-ceo-khan-academy-google-microsoft-ted-ets-higher-education-institute-bachelors-applied-ai-gen-z-college-upskilling/
New affordable housing policy
Don’t forget this market insanity started around COVID, and has basically been succeeded by consequent crises with few dips.
The typical release cadence of PC components is around 4-6 years, which requires new motherboard, CPU, and RAM.
Adding in the GPU basically results in a new build, and that’s being generous assuming no upgrades/changes to other parts like PSU and storage.
My take is that a lot of these people wishing to upgrade are those who have simply been holding out since 2020 or earlier. This seems to vaguely match up with the Steam Hardware results, with a fair number of people still using RTX 3000 series or RX 6000 series, of which even the top end cards are starting to become par/outperformed by their modern mid level counterparts.
Docker’s main advantage is just being more well known and hence more supported as a default option.
Even then, I feel that this availability of docker compose files is an illusion, due to their verbosity and limitations inherent to docker. Less granular control of permissions, clunkiness in updating images, and multi container stacks feeling like an afterthought.
In pretty much all other ways podman feels superior. Cockpit provides a basic web gui, but quadlets are the main draw. Way easier to configure, explicitly designed for multi containers, and updating all images is a single command.
Roughly, the different ecosystems from least to most complex are:
Docker/Portainer -> Podman/Cockpit/Quadlets -> Kubernetes
The most delicious museum.
IIRC they have an entire floor dedicated to variants of the scream.
MATLAB, Julia, R, Fortran, Lua.
Imo the real point of contention is column vs row major ordering in matrices. Actually so cursed trying to move between languages.
They really tried to hype an L as a W
Because people are too poor to purchase the extra output, assuming that the AI has actually lead to better productivity.
When there isn’t enough aggregate demand to absorb extra output from productivity gains, firms will opt to increase profits by reduce their costs by cutting back on spending instead.
Actually, what he meant was a seas-fire, hence the bombing of a coastal city /s
It was too good to be true. Microsoft had to implement additional 5 hour and week usage limits to cope with demand.
The problem was, for the entire history of copilot, there was literally no way to even check these usage limits. All usage counted towards these limits, even those with a 0x multiplier that didn’t consume premium requests.
It was so bad that many people couldn’t even use the premium requests they paid for the month. The only around it was to switch to auto model routing, which would tend to route to lower quality models.