> You can combine that with all the training and inference code, and at the end of the day, a system that literally writes code ends up being smaller than the LibreOffice codebase.

You really need to compare it to the model weights though. That’s the “code”.

>You really need to compare it to the model weights though

Then you'd need to compare the education of any developer in relation to how many LOC their IDE is. That's the "code".

So yea, the analogy doesn't make a whole lot of sense.

> far less performant code than the one before it.

That worked because of rapid advancements in CPU performance. We’ve left that era.

It’s about more than performance. Code is and always has been a liability. Even with agents, you start seeing massive slowdowns with code base size.

It’s why I can nearly one shot a simple game for my kid in 20 minutes with Claude, but using it at work on our massive legacy codebase is only marginally faster than doing it by hand.

The reason it’s not useful as a measure of productivity is because it’s measure of complexity (not directly, but it’s correlated). But it tells you nothing about whether that complexity was necessary for the functionality it provides.

But given that we know the functionality of Claude Code, we can guess how much complexity should be required. We could also be wrong.

>Why does it matter?

If there’s massively more code than there needs to be that does matter to the end user because it’s harder to maintain and has more surface area for bugs and security problems. Even with agents.

The he actual experiment would never make it through IRB and even if it did it’s impossible to filter out the people who spot a fake.

Gina Perry combed through all the available evidence and interviewed some of the participants and came to the conclusion that fewer than half believed it was real.

If you read Gina Perry’s critique, her conclusions is that fewer than half of the participants thought it was real.

These were Yale students, so probably smarter than average, and the study didn’t do a very convincing job make it seem believable from what I’ve read.

When I took psychology in college I had to submit to random experiments to as part of my grade (there were alternatives but the experiments were easier). Before I’d ever heard of Milgram, if one of those studies had put me in a similar situation I would have smelled a rat immediately.

When I was in middle school the teachers created a fake “government decree” to convince us that there was a new sin tax on products kids use (as a simulation). I immediately knew it was fake as did many other students, but that didn’t stop us from playing along for fun. I talked to a few of my teachers later and they genuinely believed that we fell for it.

That's pretty fun that your teachers did that. I wish teachers attempted to immerse students in the things they're teaching about more often, rather than just reading about it in abstract through a textbook or whatever.

I had a Junior High School teacher who did a variety of immersion lessons. The problem was even a small deviation from the real world structure turns the exercise into a pretty simple game. Essentially, the results are too complex and muddy to extract overall lesson.

And social science/history/economics is about learning the standard lessons of the field (even if those lessons are themselves simplistic compared to the real world, they are a baseline of common knowledge).

I did one of these experiments around 2011, and because it was so obvious that the experiment was contrived, there was a lot of misdirection around the actual experiment, which was testing something totally different from the pretense. Like different responses to font color or something like that.

>These were Yale students, so probably smarter than average

In my experience, Ivy League students are some of the most profoundly stupid people I've ever met

> Interviewing the original participants―many of whom remain haunted to this day about what they did―and delving deep into Milgram's personal archive, she pieces together a more complex picture and much more troubling picture of these experiments than was originally presented by Milgram.

Just reading the Amazon summary, I feel like there’s a contradiction. If subjects were just trying to get it over with, yes it invalidates the study but the only troubling conclusion is that the study wasn’t scrutinized more closely.

I also don’t see why they would be “haunted” by what effectively amounts to a chore to get their $20 participation check.

She never said that none of the participants believed it was real.

> But I don't see any other feasible way for people whose heads are cryogenically stored to have bodies again, except cloning a new body for them.

Well the first step would be to understand how to undo the damage caused by freezing. We’re arguably further away from this than we are from any other part of the process. We might never be able to do this, freezing might just be too lossy.

I suppose we can postpone this problem for another 100 years.

Haha. As far as I’ve heard the frozen head companies have a pretty terrible track record. The odds of them keeping the heads frozen continuously for 100 years are not good.

I spent the entire trip on the Acela Express first class (work was paying) from NYC to Boston talking to an absolutely fascinating man headed to his 60th MIT reunion.

I spent the entire trip (including a 4 hour delay where we didn’t move) in the cheap seats from Atlanta to New Orleans smelling the farts of someone with serious GI issues while a college kid walked up and down the aisle spraying axe body spray to drown out the smell.

There’s nothing wrong with doing your own experiments as long as you understand your limitations. But that’s not what people mean when they say they “did their own research”.

They mean that they went online and found blogs and YouTube videos that agree with whatever crackpot view they already held.

The issue with picking people and organizations to trust (which you absolutely should do) is that the average person isn’t even able to evaluate what qualified means. And RFK jr. is the guy appointing the “qualified people” who run things. On paper many of them are qualified, but in reality they’re crackpots.

You have to dig a level deeper and understand that this set of qualified people are actually just nuts who essentially performed the scientific equivalent of a coup because their ideas couldn’t win on merit.

To be fair, when the Covid vaccine was being rushed to be approved, I didn’t 100% trust that Trump wouldn’t pressure the FDA to approve without being confident it was safe.

So my standard at the time was that I’d take it if the FDA and at least one other developed country approved it.

Here in Argentina we approved the Sputnik vaccine. It was approved only here and in Russia. And here it was approved not by the standard office (ANMAT), but by a special resolution of the Health Ministry.

We could find it in Canada too (due to distribution it wasn't super common)

You are right https://en.wikipedia.org/wiki/List_of_COVID-19_vaccine_autho... It was approved here for 3 year old kids. Perhaps that was the odd part.