Hacker Newsnew | past | comments | ask | show | jobs | submit | iabacu's commentslogin

One thing that you overlook is something that even Xi points out as an issue: right now, China needs more quality than quantity.

For example, in semiconductors, Taiwan, with a much smaller population (and arriving late in the game) beat the US, Japan and Korea. China, with 600k STEM graduates per year, is still trying to catch up.

On covid vaccines, even Russia has their own vaccine that reach mRNA levels of protection, but China is still stuck with the Covid Zero policy.

Quantity was definitively an asset for Chinese growth based on cheap labor. That source of growth is almost running dry (to the point where they are resorting to labor camps). But at some point, quantity becomes more of a liability.


>China needs more quality than quantity.

Yeah entire MIC 2025 move up value chain drama that turned into tech war with US.

TW had decades of head start over PRC serious semi efforts and TBH if it wasn't for US sanctions on Fujin Jinhu, PRC would have EUV machines 3 years ago, might be on 7nm by now, and at least Huawei pre sanctions was world class in semi design. Issue is PRC needs to pursue indigenous quality now which is burden but also opportunity. TW will rely on foreign equipment forever, even though they have monopoly on some supply chains (which PRC is replicating). Shit hits the fan, TW semi can't work without foreign tech, PRC can at least cobble together indigenous 28nm and soon 14nm.

Flip side of US PRC containment policy is seriously taking MIC 2025, i.e. PRC has gap of 300k in IC talent (need 720k according to white paper), and integrated circuit major was only upgraded to first-level discipline in 2018 due to tech war. Previously private companies happy with using western equipment but now they're pursuing entire indigenous supply chain. Something TW can't do. PRC needs to master quantity (in every sector and subsector) and quality, which is much larger task but also large growth potential. Transition is in effect, entire soft tech crack down to redirect talent into hard tech. Excess talent isn't be absorbed in many fields but semi not one of them. But really apart from semi, turbojet/aviation, very high end machinery, PRC is closing gaps everywhere. It's matter of time after finally US tensions finally kicked PRC industrial policy into right priority.

>On covid vaccines, even Russia has their own vaccine that reach mRNA levels of protection, but China is still stuck with the Covid Zero policy.

Because mRNA vaccines makes very little difference. That's just triumphant mRNA propaganda wank. Traditional vax protects marginally less well for severe cases. mRNA virtue was preventing spread, which it doesn't post delta. Reality is Chinese vax is effectively as good as mRNA after booster, but covid zero is simply much better. Extrapolate current TW covid stats after they've capitulated to live with covid to PRC population and you're looking at millions of death even with mRNA. Zero Covid is power move by PRC who has strong state capacity to maintain zero covid (so far), relying on vax is capituation for those who systemically can't. The idea that mRNA would solve PRC covid problems is naive.

>Quantity was definitively an asset for Chinese growth based on cheap labor. That source of growth is almost running dry (to the point where they are resorting to labor camps). But at some point, quantity becomes more of a liability.

PRC labour is not cheap anymore. But PRC is improving both in quality and quantity as seen in export figures and value capture. They're moving up the industrial chain in many sectors. Remember meme that PRC only got $3 per iphone 3G?, it grew to $100 by iPhone X. Meanwhile anyone in manufacturing including Chinese will tell in you labour costs in PRC is becoming uncompetitive. South East Asia is much cheaper, but PRC efficient logistics still comes out ahead.

>resorting to labor camps

Alleged labour camps are poverty alleviation efforts as part of long running rural labour transfer programs. Even "coerced" labour studies by propagandist (Zenz) report payment is substantially higher than the 600M bottom tier Han living on 1000RMB a month. Most companies don't want to hire "cheap" Uyghur workers who has no language proficiency when there are plenty of Han willing to do the job for LESS. They're pressured by central gov take in Uyghurs workers for again, poverty alleviation (and yeah some Uyghurs being pressured to join these PAID programs along with reeducation). There's really no shortage of labours for the kind of low skill work they typically get thrown into. Like even the entire XJ cotton drama over Uyghurs picking cheap cotton is incongruent with fact that most of XJ cotton is mechanized and US sanctioned John Deere from selling XJ cotton pickers because no one in XPCC wants to spend the money to pay Uyghurs to pick cotton when machines are cheaper. Sidenote: ask why US refuse to sell machines that would alleviate alleged coerced cotton labour. Almost like they want PRC to continue subsidizing these manual labour jobs to perpetuate narrative of slave Uyghurs handing picking cotton. Anyway, XJ security architecture cost TRILLIONS with massive subsidies, even CCP chattle slaved every Uyghur it wouldn't be profitable. That's why they're focusing on poverty alleviation, with spectrum of reward/pressure/coersion/ strategies to get Uyghurs to particpate in uptraining, because integrating/improving XJ workforce and participation in broader PRC economy is long term better than maintaining securitization. You don't do that with the most inefficient labour camps, you do that by showing them (even if forcefully) that making life changing amounts of money is better than being impoverished coping on religion.


“Pay no attention to this concentration camp - really it’s a kindness being done to the inmates”


Pay as much attention as you want, kindness has nothing to do with Uyghurs being 0.2% of population and not the solution to PRC labour probems.


You missed that one by a mile bud. You’re justifying forced work camps. You’re in the wrong.


I'm point out CCP strategy on poverty alleviation and sinicization to integrate restive regions and meet development/security goals. Not interested in interrogating morality debates, right/wrong reserved for whether it works or not.


An animal sample from the wet market that was shut down. (I spooned they were all bagged).

If they were all destroyed, then go get a sample from their natural reservoir (wherever those animals came from)


This is not about children, though, it's about investors.

If Nike makes a material claim to investors (e.g. that children labor is not used), but the claim is revealed to be knowingly false, then that's securities fraud.


If the over-engineering is in the wrong dimension (one that the project doesn't need), then the cost will be double: adding the over-engineering and then detangling from it.

The problem is there are too many dimensions a given project can be over-engineered if the future is uncertain. So even an educated guess has a good chance of being wrong.


Even if the guess is reasonable and is eventually true, like "we'll need to scale, so might as well prepare for 100x capacity now", it's still frequently a mistake to over engineer.

The architecture to support something that is not needed tends to introduce rigidity into the codebase, adding a tax to future changes in order to maintain those features.


Why stop at scanning photos in your phone?

With lower power sensors, head-mounted devices with always-on-sensors, and whatnot, why not sample real-time hashes that can tip LEO about potential crimes happening?

Then why sacrifice recall in order to achieve high accuracy?

Err on the side of uploading more hashes. Then feed it all into a ML so it can use other data to filter out potential false positives.

Then if in a distance future, any LEO wants to investigate you for whatever reason, the set of potential hashes associated with your account will provide sufficient evidence for any court to authorize further non-hashed data access (it doesn't matter if they were all false positives)


The argument about POW energy is overblown.

There's an exponential decay of rewards of mining BTC (50% every 4 years).

Do you think the price of BTC will increase exponentially, faster than the decay over the long term? I don't think so.

This implies that the energy consumption of bitcoin mining will eventually decay exponentially too.


This is bizarre.

If the intern did steal code, the CEO only wants the project to be taken down?

Any IP agreement worth their salt would require Replit to send a formal/legal request asking the intern to destroy and return any stolen IP.

So I call bullshit on the CEO, and the intern should probably sue Replit for slander.


> should probably sue Replit for slander.

If the CEO is making the stealing part up (seems likely) the intern probably could sue Replit. Intentional copyright infringement/trade secret violations are a crime and my understanding of US law (not a lawyer) is that that makes it actionable regardless of damages.

Should he though? If he wins it seems likely he'll get nominal damages. He'll have invested a huge number of hours of his life into it. He'll be risking being on the hook for some or all of his lawyers fees depend on how the judge feels about awarding costs.

It doesn't seem likely to be worth it. Public shaming of Replit like this is a very cost effective way of punishing Replit... the legal system not so much.


Exactly. If the source code (or parts thereof) where copied, you could easily say which parts by just putting both things side by side. The interns project was open source after all.

This way you had to expose your own code, but it would easily win you the case in the court of public opinion. The only reason not to do this (that I can think of) is that the intern did not actually copy your code.


The difference between what Amjad tweets publicly and what Amjad threatens the ex-intern privately is jarring.

That makes me think that either the tweets are empty virtue signaling; or Amjad is legit worried that an intern open-source project can accidentally outcompete his company!


Last year people would jokingly say that the headline this year would be "bitcoin crashes to 30k", and that was dismissed as stupidly bullish.


Your comment appears to be the only one in Hacker News history that says, "bitcoin crashes to 30k". Maybe the phrasing was different?

It's not good when an asset class drops by 30% in 24 hours.


That was not a direct quote on anyone here or elsewhere, more like a sentiment that was around last year.

https://twitter.com/shortthebanks/status/1133735039596994560

https://twitter.com/Joe_wants_BTC/status/1316044080690929666

Dropping X% in 24h sounds bad, but without putting it in the context of the asset's historical volatility, sharpe ratio, overall portfolio construction, etc; it looks like a very superficial reading (perhaps just good enough for a headline)


It probably just means that it's not actually liquid at the values that btc people are crowing about. The real question is how much trade volume it took to drop it 30%.

Being at $60K just means that a handful of suckers bought for $60K.


It's not something I'd care to speculate over. There's a public ledger, we can find out exactly how many people bought it at every price. I don't have the inclination, but it's something that's precisely knowable.


No, you can’t. Given a transaction on the btc blockchain, how would you determine if actually changed hands (as opposed to internal/self-transfer), and if so, at what valuation?

Trades on sidechains or on centralized exchanges also don’t register.


The exchanges do produce a "tape" that is analyzed by various people but you're basically taking their word that it's not fraudulent.


Right.

> There's a public ledger

Nothing to do with this.

> we can find out exactly how many people bought it at every price.

None of them afaik will tell you how many people. Trade data tends to be anonymous and only give you price, amount, and side for each trade. There’s no way to differentiate between 2 people trading back and forth 1 million times vs 2 million people trading once each.


Don’t take it too literally, the point is YTD is still positive and 1YR charts are ridiculously in the green, beyond any bull’s dream not too long ago


You can always construct an arbitrary timeline where the charts are green in order to support that argument. The issue with the huge drop is the insane volatility, there are a lot of people who are freaking out because they bought at 40k+.


There were plenty that predicted over $100K by now.


I mean doge is more direct - 35 cents after a major crash. It only just got lifted from 8 cents territory, where a crash was feared


It's amazing to me that the real answer on how to make renewable energy economical (while avoiding the ecological toxicity of building utility-scale batteries) are getting downvoted.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: