According to the memo, metrics such as AI adoption rates, token consumption, code changes, and code reviews become "deeply flawed" when they are disconnected from actual results. Turning them into targets, Amazon says, encourages wasteful behavior that clashes with its emphasis on "frugality and simplicity" and can raise both compute bills and software complexity.

It's almost as if they're just now realizing Goodhart's Law exists.

And short-lived. More and more workplaces are charging the actual cost of the charge. We should expect this to continue as more people get EVs and the cost of subsidizing the fueling of commuter vehicles gets more public visibility and the value is required to be imputed/taxed as employee income.

First, the Maverick IS a truck. It's a SMALL pickup truck. You probably want a MEDIUM pickup trick, like the lightning. Or maybe one that doesn't have a back seat.

Ford says:

1. It will have a similar footprint to the Maverick
2. It will have the same passenger space of a RAV4

Ford hasn't said:
1. Towing capability
2. Bed size

You missed the point.

A friend of mine made a tea-timer app for the iPhone. Where you enter what tea you are making and it sets the correct number of minutes. Today, people would ask ChatGPT before they go on the App Store to search for something like that. That's the point. Not what YOU would do, but what the general public does. And for Jane Doe, ChatGPT has replaced a lot of these specialty apps.

Skills are great until they aren't. The problem is exactly what you say: You can literally create a skill by telling the LLM to create a skill. Which is... why the f do I need to make a skill if the AI already knows how to do it?

Properly created skills, that include reference material, specific instructions, and expert knowledge the LLM doesn't have or that's buried somewhere under 10 layers of reddit posts is genuinely useful. But most skills aren't. I'm constantly underwhelmed by hyped-up skills that, when I look at them, turn out to be bland, basic instructions.

1. Climate Change: This is decided science. The only reason we're barreling toward climate disaster is the people in political power. We were making progress on this front for decades prior to 2016.

2. Tobacco: You mean, the science came out and contradicted the lies of tobacco company, there's been a 70% reduction in smoking in adults and an >80% reduction with youth? Vaping is still an issue, but it's significantly less harmful than smoking.

3. Purdue Pharma: Purdue pharma was officially dissolved in May of this year after being successfully sued for >$7B. It's now comparably difficult to obtain opioids, particularly in the excessive amounts that fed the start of the crisis.

4. Boeing: What about them?

Pro-Tip: Life gets better when you acknowledge the improvements being made by the hard-working people around you and don't just focus on the headlines.

Actually, it DID work for climate change and smoking.

We made absolutely MASSIVE strides in reducing the risk of catastrophic climate change leading up to the early 2000s and then the US GOP started actively fighting against the science. Looking at California alone, they went from a huge amount of smog days per year to them being rare. (https://www.aqmd.gov/home/research/publications/50-years-of-progress)

And the decided science empowered social change to MASSIVELY reduce the incidence of smoking in America. (https://www.lung.org/research/trends-in-lung-disease/tobacco-trends-brief/overall-smoking-trends)

People will not make MASSIVE, generational change without sufficient reason to believe that that change is necessary.

In June, over $500,000 worth of bourbon was stolen out from under the nose of a warehouse in Phildelphia. The first paragraph:

Later in the story, the explanation for how this happened is nearly identical to these tech thefts:

Newman said investigators believe the theft may be tied to a "complex ring" targeting high-value consumer products. The perpetrators appeared to impersonate an existing trucking company, using a company name, phone number and other details to gain trust and secure the load, he said.

"They basically try and look like, sound like ⦠a legitimate trucking company," Newman said. "They spoof the company name, they spoof the trucking dispatcher."

In the company's news release, it said the theft appeared to have been carried out with knowledge of "logistics operations and product movement schedules," suggesting planning and coordination.

35% of University of California students receive the Pell Grant -- federal funding for the lowest income undergraduates. The UC population isn't at risk of losing "mommy and daddy" endowment donations. Discipline can work well-- but it needs to be WIDELY ADVERTISED discipline so as to invoke a strong deterrent effect.

You're right about the risks of just expelling, though. All public universities are exceedingly concerned about student "retention" and are very slow to actually dis-enroll a student. There's just not much back-stop funding from research grants thanks to this federal administration's new policies.

That's very idealist, but not in line with the pragmatic realities of education.

1. Most people want to HAVE learned, but are unwilling to go through the process of studying and repetition to actually learn.
2. We set up massive educational system to ENSURE that people learn. Those systems are held accountable when their students DON'T learn.
3. Checkpoints are set up throughout a curriculum to force students to learn if not for their own sake, then for the sake of a grade that will influence whether to progress in the educational system.
4. Remove those checkpoints and more students will fail. And then the educational system fail. And then there's no more advanced education.

Congrats -- Welcome back to serfdom.

I wouldn't be surprised if most of the complaints are from ai bots and not real people.

There's nothing wrong with having a community were AI content is prohibited. In the future, such areas will probably be significantly more valuable anyway.

Some social media site is going to restrict all AI and find that many people would prefer such a community.

This doesn't mean you can't find AI slop online - there's tons of places. Why not have a few that are more selective?

These are not children. These are students at a top-tier research university. They will become doctors, engineers, and career mathematicians.

Their job is to learn.

Do consider that we're talking about math at a top-tier research university.

Math classes at a University of California campus will range between 100 and 350 students. One-on-one discussions are not a functional realistic option.

Homework is graded to force students to practice because they wouldn't practice otherwise.

That's the difficulty of education. EVERYONE recognizes the need for it and almost everyone is grateful for having been educated, but 98% of the people receiving education would rather be doing something else.

Very few people are driven by the intrinsic rewards of education.