anonylizard

TSMC is banned from exporting high-end chips to China. The Taiwanese want to be an US ally, and not end up destroyed like Hong-Kong, hence they'll comply with US sanctions.

China is still far behind the US in high-end chips. Sure they are catching up. But China is on borrowed time due to population collapse. China will run out of young engineers far more quickly than the US will.

The first mover can have overwhelming advantages. If the US gets to AI first, dominates the global AI market, China facing demographic collapse, economic decline, will lose the ability to compete for supremacy

“ I think Taiwan needs to take one for the team and commit to reunification with China”

That approach was already tried against Hitler, allowing Germany to reunify with Austria (Austrians were even willing), and then taking Sudetenland.

If you reward aggression, you will only get more aggression.

Are you insane? China can only achieve that by forcibly conquering Taiwan, with minimum hundreds of thousands of casualties.

China can't win a peaceful competition against the US in the long run, because China is undergoing total demographic collapse (Marriage/birth rates declining 10% yoy every year).

China's plan is to sprint for the last stretch, while its population can still work.

As for the US, it is really held together by the massive wealth from its tech and financial dominance. Given the enormous social fractures today (please don't forget that Trump was nearly assassinated 1 month ago), an structural economic collapse will mean civil war.

You completely do not mention, that China is MILITARY THREAT.

1. China takes Taiwan

2. China dominates all chip manufacturing

3. China dominates the AI industry

4. China overthrows the US tech industry supremacy by dominating global markets

5. US economy collapses (It is not exaggerating to say the tech industry is the lifeblood of US economic growth, just look at the stock market)

The philosophical argument is actually irrelevant, since China isn't trying to export its ideology that much beyond some alliances of convenience. China never tried to win that argument either, it just blocked off the internet with its firewall, and focused on building its economy.

China is first and foremost a military threat, then an economic threat. Hence why winning AI is so pivotal (AI drones will dominate battlefields and jobs).

That's inaccurate, because traditional sentiment analysis, or rather the entire NLP ecosystem, is a very niche and underoptimized space.

Its not comparing C against Javascript, its comparing Ada against Javascript. Ada is not going to be any faster than javascript because its too niche and therefore underoptimised.

The theoretical minimum computation required by LLMs is far higher than traditional simple NLP algorithms. But the practical computation cost of LLMs will soon be cheaper, because LLMs get so much investment and use, there's massive full-stack optimizations all the way from the GPU to the end libraries.

Because LLMs WILL dominate all NLP use cases, whether you like it or not.

Its like the linux of operating systems. Sure you can handwrite up some custom OS more specialized for a purpose. But its much easier to just use linux, which everyone understands on a basic level and is extremely robust, and modifying it slightly for the end goal.

And saying "Traditional sentiment analysis" tools are "Battle tested" is laughable. LLMs in the past year alone, probably has 1000x the cumulative usage of all sentiment analysis tools in history.

LLMs get 100 billion + each year in research, improvements, engineering, optimisations.

LLMs keep rapidly improving year to year in capabilities. Sonnet 3.5 already obliterates the original GPT-4 in every aspect.

LLMs keep getting cheaper year to year. Gemini flash is like 100x cheaper than the original GPT3.5.

You can onboard any person who can write python, to start using LLMs to perform language analysis in a day. Versus weeks to use these traditional tools.

Nearly all NLP tasks will be standardised to use LLMs as the baseline default tool. Sure there'll be some short term degradations in some specific aspect, but there's no stopping the tide.

By the way, traditional ML-based translation is also pretty much dead and replaced by LLMs. I've been seeing an explosion in fan-translations done by say Sonnet 3.5, the improvement in fluency and accuracy is just radical and extreme, I often don't even notice the AI-translation anymore.

LLMs massively compound the advantage of existing popular languages, namely python. Any new learner will find it infinitely easier to use sonnet 3.5 to overcome the so called '2 language barrier' for python, while the lacking data for Julia becomes the real barrier.

This issue will remain until LLMs get so smart they can maybe self-iterate and train on a given language. By then though, we'd likely get languages designed and optimized for LLMs.

Ehhhh.... Why can't the government keep the road neat? Isn't that the government's job?

The problem with including a stipulation in the contract, for the private company to keep it neat, is its hard to objectively monitor (Unlike say the money paid for advertising), which will therefore 100% become a way to extort the private companies.

The optimal solution will just be 'pay off the government inspectors & spend as little on actually keeping it neat as possible'.

People need to understand a government policies is like a software program, its not trivial to just 'add a feature'. An extremely corrupt government, is like hardware that flips bits randomly all the time, so please don't make it do even more things.

Are you people being deliberately obtuse? A presidential candidate is equivalent to some random people? Like it was 2cm away from civil war.

Are some random gang shootings that killed 10 people worse than JFK's killing?

The university presidents got 'resigned' due to failing the congress questioning.

If you are working for public institutions, then it probably matters immensely. But if you answer only to shareholders, then its more theatre.

The head of US secret service just got 'resigned' after being double body slammed by Republicans and Democrats in congress on live TV. Still that is getting off lightly for a catastrophic failure 10x worse than crowdstrike.

Society was able to move to mass WFH on a global scale in a single month during Covid, thanks to the highly centralized and efficient cloud infrastructure. That could have easily saved tens of millions of lives (Imagine the Spanish flu with mass air travel, no vaccines, no strain-weakening)

These small 'downages' basically never cause serious issue. Your solutions are just alarmist and extremely costly (though they will provide developer employment...).

Like how do you plan on competing against multimodals, which keep getting cheaper and clearly can do audio->text? Or existing incumbents like Deepgram? Or just the generic APIs provided by the big clouds.

AI allows people to 'express themselves in unique and interesting ways' far more than VR does. Text and image generation allow people to create worlds completely tailored to themselves, this time unlimited by the need for other people to play along.

So AI competes with VR both for users (The type willing to pay for cutting edge consumer tech experiences) and investment (AI is the big thing, VR is so 2022), and labour (How many researchers want to work in VR over AI?)

It allows for looting the national treasury today.

All large construction projects in Egypt go to the military first. Who then subcontracts it out, essentially paying the generals their cut. Hence why Egypt is building their new capital city, to placate the military elite.

The new capital is also to riot proof their capital. Large dense cities are hard to defend with armies. Hence Myanmar doing the same, even China is building a mini capital to the south of their existing capital.

That being said, this big museum does make sense. Egypt is tourism dependent, and it needs star attractions beyond the scammer infested pyramids. Egypt's reputation amongst tourists has taken a horrible dent due to the extreme concentration of street hagglers, and it needs all the help it can get.

Yeah bro, why have cars, we have legs! Why have teachers? We have books!

In reality, modern society has plenty, plenty, plenty of obstacles for every age group. Our priority should be eliminating them, not worrying about having too few of them.

The sheer mountain of existing knowledge that we have built up, is sufficient cause to use AI assisted teaching. Children in the future, will not only learn physics, chemistry, computers, history, language etc, but will also learn carpentry, cooking, personal finances etc, all helped out by the drastic efficiency leap from 1-1 personalized AI tutors.

The huge solar power plants in China are in western China, aka deserts and frozen desert mountains, aka you can't build data centers there either. And shipping that power across thousands of KMs is as hard for China as it is for the US.

Hence China is still building coal plants, and won't really stop. Especially if it means the future of AI is being bottlenecked by it.

Why is an employee obliged to 'make things work', if those very same things are going to threaten their job?

I'm also not certain how this is sabotage. If executives treat developers like interchangable factory line workers, then its important there is a like-for-like comparison between the contractor workers and the onshore ones. Ie working with minimal guidance and high quality standards.

Anime already cuts corners by using cheap 3d models in place of 2d hand drawn objects.

The more rigid the object, the better this works. 3d cars look better than 2d cars, even in an otherwise 2d show. Mechs look a bit worse. And human characters look horrible.

Yet anime studios still do it. Including for critical highlight scenes like dance scenes (Check out Love live dance scens), because it is so, so hard to draw humans dancing.

So if anime studios are willing to do something, that looks obviously bad, as a widespread practice. There'll be 0 barriers to AI inbetweening adaptation, which would likely look BETTER than human inbetweening within a year of release.

AI anime art has already wiped out the lower-end of patreon artists, and is heavily impacting the mid-tier. Because AI has gotten more technically proficient than the average mid-tier artist. Pretty much only the higher-end can hold their heads above water. Or they have to transition to drawing comics with storylines, instead of just simple images.

Wow, this looks like the beginning of the end for in-between animation.

A large chunk of an anime's budget goes to in-betweening. Essentially human interpolation of graphics between two key frames (Its usually like 6-12 inbetweens per key frame). People hate this job, and it is highly unproductive, so generally outsourced by Japan to other countries.

Western animation decided to abandon it altogether, and move first to flash, then to 3d animation. But in retrospect that was a mistake, as it lost so much of the creative flexibility of 2d animation. Anime today is substantially bigger than western animation as a result. Crunchyroll has 13 million subscribers.

AI will solve the problems with 2d animation. Something like SORA fine-tuned on anime-keyframe data like this. Can probably easily solve in-betweening. Then the 2d animation workflow will dominate 3d. Its so much easier to just draw a beginning and end key-frame, then have the AI fill it in. Than to model and rig and render a entire scene.

No they won't lose a lot of users, what's the competition or alternative? ChatGPT? Even more AI?

Youtube did like multiple redesigns, each time with massive complaints, clearly had no negative effect on youtube.

Reddit did that huge redesign, and lost like 50% of users initially. Their profits keep going up because they now control the ad distribution and can protect their valuable data from scrapers.

A UI/UX degradation simply isn't enough to turn away free users. You have to damage the fundamental value proposition like say Tumblr banning adult content.

This, starting out at a startup. When you later go onto a big company. They'd be surprised by your ability to build a product end to end, from design to implementation to infrastructure.

The only part that's missing is handling the enterprise processes, multiple stakeholders, chains of approvals etc. But those are easier to learn than building something from scratch.

Like have you checked the license of Llama3? Who told you it was free for all commercial use? Most open source licenses by big companies these days, are anti-big-tech. Aka anyone can use them but big tech.

"If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee’s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights."

Fine tuned LLMs can play chess at grandmaster levels.

So its clear that there is in fact 'deeper patterns to chess' that allow one to play very well, without any search required (Since LLMs cannot search). Its just that those patterns are probably rather different to human understood ones.

Certain people, discovered how easy it is, to change entire institutions, by applying targeted pressure towards individual managers of those institutions.

No manager dares to refuse these initiatives because they'll be instantly named by the activist, with the help of journalist friends, and ensure they lose their job.

Hence by going one at a time, a small group of activists can change far larger organisations.

Of course, every action has a reaction. Enough resentment has built up, that made anti-woke a legitimate position to take in the public eye, with no longer much risk of losing the job. This makes all the previous tactics lose their power.

Also, the 'anti-woke' crowd is learning these tactics too. From congressional grilling of university presidents, to gamers explicitly documenting the DEI consultancy involvements, its all targeting and absolutely destroying individuals to cause a chilling effect on the populace at large.

People like you, is why Europe passed on global leadership to the USA.

Ukraine can be burning on your front door, and you still think EU is safe. Guess what, EU is safe only because of NATO. Russia has attacked basically every neighboring country in Europe that is not in NATO.

Nukes don't work against salami slicing. Russia takes Estonia tomorrow, are you going to throw a nuke against Russia? No, you'll won't risk annihilation for Estonia, so you'll ignore it. So the next day Russia takes Latvia. And so on so forth. The only way of fighting salami slicing is applying proportional response, which requires a conventional army.

Also, Ukraine living standards isn't very high right now, regardless what their PPP is, and getting conquered by Russia, means you get used as battle-slaved for the next conquest. This snowball tactic was innovated by the Mongols, using the conquered people as cannon fodder for the next war.

Russia and China cannot deliver more firepower per USD. Any such impression is because Russia/China only fights near-shore wars, rather than the expeditionary wars of the US.

Because the US fights global wars, it focuses on quality over quantity, as logistic costs start to dominate the production costs of military power.

If China tried to fight some Iraq war, it'll be just as expensive as the US effort. Even soldiers have to be paid far more in offensive expeditionary wars than defensive homeland wars, just to maintain morale.

One exception is mercenary groups like Wagner, they are actually extremely cost effective. That's because of much more flexible management structure, using private enterprise structures instead of rigid formal militaries, works 100 times better for chaotic low intensity warfare.

Russia is not manufacturing more weapons than its GDP suggests. Its simply spending more money than the west donates to Ukraine (Because Russia is fighting for Putin's survival, while its not #1 priority in the west), and importing everything from China using that money.

Russia is actually precisely the example why GDP matters. Russia's position before the war was an overwhelming advantage, pinching Ukraine from three directions in a huge concave, in a flat plains basically without geographical barriers, against a divided, poor, and old (unlike say Afghanistan which has endless 18 year old men) country in Ukraine.

Yet it has fought the war into a extremely expensive stalemate, with draft imposed, large casualties, and basically total loss of international prestige.

All that, because its low GDP meant no high end equipment, no advanced logistical capabilities, no sophisticated information advantage. While, the US, by Elon Musk simply donating some starlinks (A definitive high GDP product), made all Russian attempts at disrupting Ukrainian command chains useless.

Manufacturing is not a magical lamp that craps out guns.

It is merely a part of a value chain. Namely, minerals and petrochemicals have to be produced upstream. Financial GDP captures the ability to produce or purchase the upstream resources, manufacturing alone does not.

Now manufacturing capacity is illiquid, and it takes time to convert financial power to manufacturing capacity, see the Ukraine ammunition crisis. While manufacturing capacity is more war-ready.

However, China hasn't invaded Ukraine yet, because US choking off the sea lanes will be catastrophic for China's economy and offensive war capacity. Without trade, the factories have no resource inputs, and no markets to export to.

So in the end, financial power does slightly underestimate the hard-power benefits of manufacturing, but GDP (USD) should still be used as the primary determinant of hard power, not amount of factories built.

The entire article can basically summarized in one sentence

"Using PPP instead of USD to measure GDP, significantly shrinks the gap between the EU and the US"

Now, using USD has its problems, but PPP's primary purpose is living standard comparisons, and doesn't work as a national strength comparator.

When in a war, you are buying ammunition and weaponry with USD, not PPP. When there's a energy shock, you are importing oil and gas with USD, not PPP. When you are in a AI race, you are buying GPUs with USD, not PPP.

There'll be some people who argue that 'only living standards' matter. If the two simultaneous hot modern wars in Ukraine, Gaza, and soon Taiwan can't convince that hard power matters, then don't complain when the draft officer knocks on your door.