PlatoIsADisease

>It’s clear that the era where any one country has global hegemony should end.

Unipolar worlds are safer than Bipolar. Multipolar is extremely dangerous.

I imagine you didn't know that more people will be killed if the US doesn't have hegemony.

In classic Apple fashion, they fooled people into thinking an integrated GPU is the same as Nvidia.

Gosh I wish I could hire their marketing company.

Unified Memory and Integrated GPU.

Apple is amazing at marketing to make 1990s technology sound cutting edge. I'm sure they change something for plausible deniability, as a nominalist, not even 2 of the same computers are the same.

The wild part is that these are awful and not usable.

Both my fortune 20 company and my buddy got these for LLMs... and the champion/my buddy had the look of shame when it wasnt usable.

Ask any AI, they say Apple has the best marketing of any company in history.

All this tracks. Anyone else getting bombarded with WSJ youtube videos about Apple manufacturing?

Have you looked at Motorola? I'm not sure they have all of those features, but me and you think similarly and when I did research, I ended up choosing their $130 phone for my contractors.

But I main the $900 pixel.

They are so similar its weird, but Motorola was slow with snapchat and the keyboard some time.

To be fair, no one should be using Ubuntu. They are the free CD people from the 2000s. They are the Apple of Linux, marketing wins, but low quality.

They used outdated linux (Debian-family) because its lower cost to maintain.

All around, never use debian-family outside servers. Fedora is the future. Maybe OpenSUSE too. (Note these are not Arch or related to Arch)

After using OpenClaw for 1 week, I'm so extremely bullish.

Buy buy buy buy.

We don't even have enough data centers.

Question for you, that 13t/s, is that pretty solid even with high context/tokens?

I know Apple marketing says 'look at our 20t/s' but they sent less than 40 tokens.

This is not 'running it on our own computers'.

But I like the option to give my data to a rando rather than one of the big 5 US companies that can get sued. At least the rando probably has no idea what to do with 10M of my customer's IP.

Actually... thank you for the links. Unironically.

Go free stuff! But... no one is running 400B models on their computers.

You are just giving them data instead. Its not like China is known to protect IP. Your data is going to be used against you, and we cant use western laws to keep it safe.

OpenAI has done more for society than all academia over the same timespan.

You could probably say the same about various medical devices/medicines. The internet. Chemistry/plastics.

Sociology/psychology is historically unfalsifiable, and mostly useless, those are truly detriments.

quantum physics might be useful, but I think even there, industry will do it better.

If I was in my early 20s, this would be mad respect.

Now that I'm in my 30s and I know PhDs.... They are basically nepo babies who were not good enough for industry.

There must be something missing if they are religious though.

Like some sort of critical thinking isnt there.

Academia is basically outdated, or needs incredible reform.

Industry and youtubers are making significant scientific progress. (I'm mostly joking about youtubers, but it does happen)

I think Academia is where B/C-list performers pretend they are A-list.

Well I wasn't using your homegrown solution, but now I'm using openclaw and it 'hijacks' my computer to do whatever I tell it.

Heck, my 6 year old kid has been using it.

My 6 year old kid isnt writing code in langchain.

Couldn't you make it just as simple with Linux? You are just doing more.

t. Someone with a dozen Apache servers that I only touch when I hit end of life.

you give an LLM control of your computer.

Yesterday I told it to make a website and it opened the browser, did a bunch of steps, (I did have to authenticate). But then it connected some html on my computer with a server with google sheets.

Consider its a massive security risk. You are giving it full access to everything your computer can do. (Potentially, you can limit stuff)

Thank you for saying this. I read this and was like: wtf?

Love agents, but the security risk is insane.

I don't think there is a correct answer here. Induction is great for short term and easily covering a wide range of subjects even if it only covers a tiny part of that subject. Deduction is great for long term progress, but only in narrow parts of the subject, but has a near universal application when found.

Use both?

Not sure if this is what you are looking for, but there is a bit of science on this. Here is a reddit thread discussing it.

https://www.reddit.com/r/psychoanalysis/comments/1r6h9h5/any...

I specifically liked the paper:

https://www.apa.org/pubs/journals/releases/xge-xge0001799.pd... https://www.nytimes.com/2025/06/30/well/mind/cool-people-tra...

I think cool people make more money, have more opportunities, probably have more fun. But don't get me wrong, the epicurean hedonist in me sometimes wants to just chill and eat bread and water.

Early in my career, I'd go to work and ask people 'do anything last night?' and we would both basically say: Not really.

I decided I did not want to be boring. I decided to spend an hour at least on something I found interesting or economically useful. I started a company, would learn programming(now I'm a pro programmer), I learned a few different arts (great for relating to a different set of people when you explain you draw, paint, sew, and crochet)...

Whatever the case, I think there was economic benefits to 'not being boring'. However you really need to push yourself, its way easier to veg out on the couch to fiction. I think caffeine and weed helped me initially, now its just my normal lifestyle.

Well it got all 10 incorrect when I asked for top 10 catchphrases from a character in Plato's books. It confused the baddie for Socrates.

As someone with a 3060, I can attest that there are really really good 7-9B models. I still use berkeley-nest/Starling-LM-7B-alpha and that model is a few years old.

If we are going for accuracy, the question should be asked multiple times on multiple models and see if there is agreement.

But I do think once you hit 80B, you can struggle to see the difference between SOTA.

That said, GPT4.5 was the GOAT. I can't imagine how expensive that one was to run.

Induction vs Deduction.

Grounded theory is probabilistically correct. Deduction if correct, is actual reality.

Don't get me wrong, I want to love induction, I have William James of Pragmatism on my wall... but the problems with induction hurt me to my core. I know deduction has problems too, but the Platonic Realist in me loves the idea of magic truths.

I worry Colonialism will come back again. Or at least in a different form.

Most International Relations practitioners are followers of (Systems) Realism. You might find some minor power with Idealists/Instituitonalists, but they only get that privilege by being under the umbrella of a great power.

Colonialism was not some greedy merchant/state thing, it was an Arms Race. It follows the inevitable forces produced by anarchy, there are no police to call so power is the greatest form of security. It causes a Tragedy of the Commons situation in the form of an Arms Race.

After Colonialism, we had essentially client states, which seems similarly brutal.

What models are you running locally? Just curious.

I am mostly restricted to 7-9B. I still like ancient early llama because its pretty unrestricted without having to use an abliteration.

Only using my historical experience and not Gemini 3.1 Pro, I think we see benchmark chasing then a grand release of a model that gets press attention...

Then a few days later, the model/settings are degraded to save money. Then this gets repeated until the last day before the release of the new model.

If we are benchmaxing this works well because its only being tested early on during the life cycle. By middle of the cycle, people are testing other models. By the end, people are not testing them, and if they did it would barely shake the last months of data.

Buddy, my wife vibe coded an app that solves her PTO tracking problem in 1 shot.

Her payroll company was going to charge her an extra $200 a month.

>24 tokens/second

this is marketing not reality.

Get a few lines of code and it becomes unusable.