Fascinating, thank you for posting.

Why did you choose to start supplementing with lithium?

Thanks for sharing, that's fascinating. I've only started using ollama within the last month, and I'm really, really regretting I didn't load up my MBP with more ram. When I bought it, 36gb seemed like more than I would probably need..

What models are you using for coding?

What CPU? How well does the CPU + 128gb ram run the models we're talking about?

I watch dogs (primarily overnight--most for 3-7 days but some 1 day and some >7d) via Rover. I make around $1500/month (pre-1099) and after their ~20% cut (of which most people give back to me in tips).

I WFH so the largely passive income is nice. I wouldn't have found as many people w/o a platform to do the heavy lifting for me in finding new dogs.

I am not advocating that we need to have these sorts of things in the market, but it does make for nice extra cash. YMMV.

Imagining a Beowulf cluster of firsts.

For whatever reason people are confusing "made in America (USA)" with "made by American (USA) companies".

While some people costs are obviously more, the benefit savings operationally could make up for most if not all of the cost leading to a "made in America" product that is same ... price wise.

However, that assumes a lot. While it assumes that all things can be done in America (likely by foreign owned manufacturing/assembling residing in the USA), for me, the bigger variable is that outside countries play by "different rules". For example a "company" in a foreign country may be slight extension of the country itself (owned by the nation/state). And therefore, from an operational perspective, they sort of play outside the "rules". Not saying that American companies always play by the rules... but sometimes we make a bad assumptions comparing "our country's" behavior to "other countries" behavior.

I once worked for a startup with patented tech. And another company leveraged that tech. We sued. Than company sold out to one of those "other countries" where the new parent company was effectively an arm of the country. So... now our lawsuit is against a country. That is, we lose.

Which is to say, if a "made in America" product becomes competitive, other places outside those interests, could "fib" (or whatever) in order to undercut any perceived advantage.

This happens all the time in some of our markets.... but as (in the past decades) we moved to a pure consumer base and service only economy, it wasn't an issue. Since everything is made in China (for example), there's no issue (as far as China is concerned).

Covid taught us how fragile being completely dependent on others is. However, that's not even talking to the vulnerability of sovereignty. I think that was the wake up call. Supply chains are one thing. Being consumed or taken over without a war, a much bigger deal.

There's a lot of variability here. The Seattle and California models are baffling to me. Totally agree that the US system is largely focused on irrelevant (or immutable) things. I have a kid in highschool now. His peer group is very, very impressive. Multiple perfect scores on the ACT every year. ~55% of the highschool is English-language learners and 60%+ is free lunch eligible (meaning poverty level or close). There's an engineering magnet program that does really, really cool stuff. The school was on lockdown 3 times last year for gang fights.

It's a tough environment, comparatively.

I have a different take! What's the point of a test that nobody ever gets a perfect score? I guarantee you that I can design a test that nobody ever aces, but it also wouldn't be worth anything. I think there needs to be a middle ground between overtesting, teaching to the test, and tests being the be all and end all of education, and the loosey goosey approach one often encounters in the US (most commonly among leftists) that thinks all testing is bad and racist and invalid and hurts kids.

You are probably assuming the wrong things given the demographics of Slashdot and those few of us who have hung on for decades at this point! I was not on the math team, but I had friends who were. I participated in both Latin and Computer Science competitions (and marching band). Our football team sucked (I know this from marching band). But yeah, I'm sure a huge amount of money was spent on the gym and fields and athletics, far beyond what was spent on supporting the best academic achievers.

You will get NO argument from me that America's obsession with sports, from the cradle to the grave, is hugely detrimental to our society and culture.

I hope that the current mess with NIL, paying college athletes (I will NOT call them "student athletes" -- what a joke) forces some or many schools to back off on their sports expenditures and focuses, but I'm not holding my breath.

Hey, I asked you your metric, I wasn't planning on nitpicking it!

Sure, just like micro-regions, individual demographic groups in the US, etc., score higher.

IMHO, US public education is amongst the very best in the world at the high end and pretty bad at the low end. The real confounding factor is that demographics are hard to escape.

I agree PISA is a reasonable standard.

So, in the context of STEM that you raised, the US scored 14 points above average, #16 in Science placement. That seems strong to me.

Math is slightly below average (-4 points), number 34. Weak.

Reading, US placed #9, seems strong again.

You mentioned immigrants being the majority of STEM students. I don't know if that is true, but, using the metric you picked of PISA scores, some of the countries that send the most immigrants to the US in STEM .... have no data available. Nothing for China, India, Pakistan, etc. India seems to have scored exceedingly poorly the last time they participated.

The chart at the bottom of the Wikipedia page that includes 2015 US State results and racial breakdowns for the Mathematics portion for the US over multiple years is fascinating, and I think, should make an impact on your thinking.

(For instance, in the 2018 math results, Asian students in the US scored 539, whites scored 503, and the US average was 478. That puts US Asian students between Hong Kong and Taiwan in position #5 and white students in between Seden and Finland at #20. Black students in the US scored 419, the level of Thailand and Uruguay, around the upper 50s in rank. Kind of changes things a bit?)

What metric would you use to compare K-12 public education systems around the world?

Both of my grandarents died from lung ailments.

One of my grandfathers worked in factories his entire life, smoked for about 20 years of his youth, and and was generally of the socioeconomic level that wasn't able to live an extremely healthy lifestyle. He developed emphysema and associated conditions, and died in his low 70s. He had been on oxygen for years at that point.

My other grandfather never smoked a day in his life, never drank, worked in academia, swam daily until sometime in his 80s, and died right around age 90--from lung cancer! One of the doctors asked him if he had any career asbestos exposure and he answered "Well I did go down into an asbestos mine in 1934..."

You never know what will get you, but dying at age 90 and being in good physical and mental shape is pretty good by me!

I strongly agree with almost everything you wrote. Taking the locale example, and using my own state as an example, most counties have their own school system, and a handful of cities have their own school systems. Even within a single county's school system, the disparities in results between school are very significant. They also track, almost 1:1, with socioeconomic level, and that in terms very closely tracks with race. I'm absolutely not making an argument that race is a cause of educational disparities, though I would not rule out any genetic impact, but race and socioeconomic level are tightly correlated almost everywhere in the United States (and really, the world).

I googled a few lists of top US states by K-12 education. There are a lot of variations, but, as you said, New Jersey (just over 50% white) shows up near the top in most, so does Massachusetts (another state with huge immigrant and non-English speaking populations, but that is still ~75% white). Many of the other habitual high achievers are small, (more) homogenous, wealthy. Even the link between spending on public education and success isn't written in stone. California, for example, spends a lot, yet gets a pretty middling result on average.

The "homogeneity" argument as you call it, and I would note that I listed 3 criteria present in most of the top countries, primarily small, ethnically and culturally homogenous, and wealthy, is just one part. All three are important factors. South Korea, Denmark, Netherlands, Belgium, Slovenia, Japan, Germany, Finland, Norway, Ireland, Singapore from one list. Most are small (Japan/Germany/South Korea definitionally), most are wealthy (I guess Slovenia would be lowest?), all are pretty ethnically and culturally homogenous, with some variations. It will be interesting to see if, for instance, Germany's results change given the massive population change in the last 10 years there.

The US has a strong K-12 public education system (not the best, but still strong) and an incredibly good showing, the best, in the list of strongest universities in the world.

When you look at most lists of the top countries in the education ranking, they are primarily small, ethnically and culturally homogenous, and wealthy. (And to be clear, not all top educational countries are all of those things!)

If you compare white and asian student achievement in the United States to elsewhere, you're going to see results very comparable to European and Asian countries. The United States, unfortunately, remains highly uneven, and our population is far more diverse (racially, linguistically, culturally, and probably IQ-wise too) than almost any other high-performing country.

Does it? It would be fascinating if AI could eventually solve problems like that, but I think the economic justification is more along the lines of "we have billions of dollars of capital available, billions of people who want to use this technology, tens of thousands of other companies working on AI technology, and we don't want to be left behind when the next big discoveries are made."

That's a really good point, and I would love to hear the GP's response! Why isn't alpha-beta pruning considered reasoning? Why isn't A* search over a generated graph considered reasoning?

I've said it before, but I truly do think many of the most vitriolic anti-LLM voices come from a place of being worried about what LLM capabilities say about the nature of human intelligence and reasoning. Perhaps the bottom line is that human intelligence is not so special as we make it out to be?

Go stands out to me as an interesting case. When I took an AI class ~2004, Go was given as an example of a seemingly intractable problem--the search space is simply so big that no existing AI-methods could effectively manage it. Many people talked about Go in almost mystical terms, as a problem that only human intelligence could approach and one that was fundamentally different from chess (now derided as solved and a mere search problem--no special human intelligence required).

It took only ~10 years and by 2015, Google's machine learning models were crushing the best humans.

I'm fascinated to see where the technology goes.