The thought they have immediately before or after is also, “I’m the next Steve Jobs.” It’s always about the ego before any product or need.

Toddler’s drawing ass-object?

Pretty much what they did in China as well. Google, back when they had a backbone, decided to leave instead, and Apple was like, “How can we censor ourselves better for you?”

Musk is the living embodiment of the Dunning-Kruger effect, and his wealth represents the exact monetary value delta of confident ignorance over intelligence.

In his memo, Zuckerberg predicted that even more difficult days could be ahead for the company, despite vowing to hold off on any future layoffs for the rest of the year.

“Given the complexity of these changes, we’ve made mistakes and will almost certainly make more,” he admitted.

“I will certainly make more mistakes that you will have to pay for, but I can promise you you will probably not be laid off for at least six months.”

I don’t suppose there is an easy to create device that creates tens or hundreds of random virtual bluetooth and wifi devices that we can leave in our vehicles to poison the data it is collecting?

Someone in MS saw the old Outlook, heard the complaints, and thought it was a great opportunity to lead a full rewrite and new product launch in order to get promoted.

That seems like basically default Moore’s Law which predicts performance will double in 18 months (at least the current iteration), though?

8x = 2^3, or three doublings (1->2->4->8).

3 doublings x 18 mo = 54 months.

Compared to 5 years x 12 mo = 60 months.

Your choices: (Yes, use Copilot) (Remind me later)

I think you’re getting piled on here too aggressively. This is literally “nostupidquestions,” so I’m assuming you asked because you’re open to explanations, not to troll or start a fight.

If that’s true, here are some reasons to be very skeptical and careful with AI (especially corporate AI; I’m neutral myself on fully local home AI), and a last line how that turns into “hate”:

• They are not as good as you think they are, which is devious. They will convince you they are “smart” but they are just statistical models. So they will confidently tell you false things and if you trust them, you will believe false things. They can’t do math, they can’t actually “think.” You are just getting a statistical approximation of grammatically plausible language from the training data.
• They train you to not think for yourself and rely on them. There is some ambiguity whether there may be complex benefits, but it’s clear how they’re being used now is harmful to learning and development, even used by people who think they’re being careful.
• They asymmetrically benefit owners versus workers. Corporate AI is being pushed so hard because owners believe it will further funnel more income to them versus workers, and they benefit from this trade even if it can’t do tasks as well as workers. That means less jobs for the workers and worse experiences for the customers.
• Huge environmental costs for all of this.
• Data centers take up huge amounts of water from communities.
• Data centers increase electrical costs to communities.
• Huge increase in consumer hardware prices from corporate AI buying all the graphics cards, CPUs, hard drives, RAM, etc, leading to pricing out many people from home computing or gaming.
• AI tools are privacy nightmares. Everything you say in moments of vulnerability will likely be used to sell you something or against you in the future.

Are there benefits? I actually think there are. But the costs are very disproportionately high. And AI has not been allowed to just grow and specific uses to be shown useful over time. It’s been shoved down our throats. With the above costs which most of us online see, and didn’t agree to, that motivates a lot of legitimate hate.

There’s an irony to Zuckerberg’s and the rest of the oligarchs’ fevered delusions about AI, and it would be pithy if it weren’t so tedious:

They wanted to discard their need for other humans, and all they discarded was their own humanity.

While planning is ongoing and details are in flux, discussions have centered on having the firms voluntarily cede the shares to the government, the people said. The returns on the investment could then be directed to public purposes, one of the people said, such as distributing a dividend payment to all American households.

I don’t believe either of these two assertions has any chance of happening.

But such an arrangement could also pose novel governance challenges, given the complications of the U.S. trying to effectively regulate something it partially owns, while also arguably increasing the incentives for a federal bailout.

Even if by some miracle the US has “shares” of the companies not paid for by tax dollars, it creates an anti-regulation incentive, forcing the public’s interest to align with the AI companies’, which is going to be worth at least whatever we would have paid to those AI companies.

Ah. They found fingerprints but lack any identifying evidence linked to a single suspect. Guess they gotta close the case!

As I understand it, unfortunately most of the memory that has already started manufacturing or been sold will not be usable, since it is specifically packaged HBM for AI servers/systems.

So once the crash happens, that is presumably the beginning of a longer re-supply ramp-up.

I think it depends on the drive sizes, whether you accept used, and if we’re comparing bare to shuckable, but yeah, for drives that I am looking for in the 20-26TB range (new, since I don’t have enough parity/redundancy to trust used drives), it seems more like 2.5-4x cost.

I don’t doubt you found a deal somewhere, but here’s what I’m seeing:

$229 for 12TB used SAS: https://diskprices.com/?locale=us&condition=new%2Cused&capacity=12-12&disk_types=external_hdd%2Cexternal_hdd25%2Cinternal_hdd%2Cinternal_hdd25%2Cinternal_sshd%2Cinternal_sas%2Cexternal_ssd%2Cinternal_ssd%2Cm2_ssd%2Cm2_nvme%2Cu2

That’s admittedly better than $26/TB, but I also had been looking at 20-24TB drives. That’s the other change - the lowest-$/TB (highest value) TB size is has decreased substantially (from around 20-24 in late 2025, to 10-12 now).

To understand the issue, I went back to a book published in 1954, 20 years before I was born: Peter Drucker’s “The Practice of Management.” Drucker explores the different roles inside every business, which I would categorize as builders, sellers and measurers.

…

Measurers are also critical to a business, but different from the other two. The best are hard to find. They work tirelessly behind the scenes, don’t seek the recognition of a front-of-house role, and ideally have a perspective independent from the rest of the organization. Drucker argues that measuring business is important, but customers are earned through building and selling. The best businesses would maximize investment in those two functions.

AI isn’t coming for builders or sellers, but it is coming for measurers. Tireless, independent, efficient and available, AI systems can now measure an organization with a level of objective detail and precision that was previously impossible even for the best employees.

It’s surprising how typical, how mundane, the logic errors CEOs make are. This guy read a book that he admits was released (presumably he means as a signal of its timelessness) 20 years before he was born. It collapses everyone into three groups. He assumes this heuristic is both correct and accurate to modern times, simply because the words have meanings that attach to modern concepts.

But builders in 1954 were making widgets and physical objects. Sellers were often in a store talking to customers who wanted to buy that object. Measurers were bean-counters, barely out of slide-rule days.

All of these types of roles have updated and transformed to the point of being totally different, and now - like with Cloudflare - the main product is information (code and knowledge, packaged as Cloudflare’s site products). Builders, sellers and measurers all just interact with that information in different ways. Each one has different efficiencies and value.

And here at least, assuming that AI is going to capture “measurers” more than others is not a smart CEO’s inspired leadership. It is a symptom of a brain that can’t adapt to the modern era.

Indeed, why would he not assume that the “builders” will be replaced more easily, as most other big tech CEOs think? Isn’t Claude Code the one success story in a sea of AI losses in the billions? Presumably his product can’t infinitely scale, so why would he keep hiring into infinity? (Those assumptions supporting firing “builders” are wrong, of course, but are as valid as his wrong assumption.)

Why would he not assume the “sellers” can be replaced - their entire job is chatting and collating information tailored to prompts as an intermediary, which is the entire purpose of LLMs? (Again, I don’t think they can be replaced, but again, it’s equally wrong in the ways he is wrong.)

The problem, as always, is that CEOs never understand exactly what is happening on the ground. They need these heuristics to give themselves confidence that they understand the whole picture. The problem is at some point, they lose their humility, their connection to the people who are doing the work, enough to care. If Prince cared enough to think harder about what these “measurers” real value is, and what AI’s real competency is, 20% of the workforce might still have a job.

That is a worst case horror story that everyone should think about, and I’m not usually an optimist, but I don’t think it’s likely.

Ultimately AI is a hype wildfire, and it will eventually run out of fuel - signs are already showing that happening as AI hyperscalers and vendors are ending investments, restricting access and raising prices to recoup unsustainable losses.

At that point, I hope we stay sane and not jump at the first discounts, and just sit tight while prices return to normal. Prices need to fall heavily before we start supporting these AI-first companies again, or else we are going to lock ourselves into that AI-inflated price dystopia.

Much more than doubled. Most high-TB drives are not in stock anywhere, and even if you find a drive, the best deals are around $26-28/TB for used drives, whereas before new deals would be $10/TB. If you’re looking for a specific new capacity, you may be paying $36-40/TB.

“Only do it in the way that makes me more money.”