>he thinks there is an 'ai bubble' that is about to 'burst'
>他觉得有个‘AI泡沫’马上就要‘破裂’了
>he thinks the price of RAM and vRAM is EVER coming down
>他觉得内存和显存的价格有朝一日会降
>他觉得有个‘AI泡沫’马上就要‘破裂’了
>he thinks there is an 'ai bubble' that is about to 'burst'
>他觉得有个‘AI泡沫’马上就要‘破裂’了
>he thinks the price of RAM and vRAM is EVER coming down
>他觉得内存和显存的价格有朝一日会降
AI is going nowhere.
AI不会消失的。
4GB RAM is the standard now.
4GB内存现在就是标配了。
>>109453437
>We can get away with charging these prices and they're still paying it... Why weren't we making them pay this much before? Goddamn we've been missing out on printing money
>我们收这么贵他们照样付……以前怎么没让他们掏这么多？靠，这些年真是错过印钞了。
>>109453437
goyim have to cope somehow
韭菜们总得找个法子自欺欺人
ban doomnigger demoralization threads
建议禁掉末日黑子发的丧气帖
>>109453437
>have many gigabutz of ram
>内存好几个G起步
>it's never totally full
>从来不会完全占满
>don't care
>我不在乎
I laugh at people who call AI a bubble.
我笑那些说AI是泡沫的人。
It's like saying there's an atomic bomb bubble.
这就跟说有原子弹泡沫一样离谱。
If humans can harness the power of God at their fingertips that could crush other nations, you can bet your ass we'll keep spending money until it happens.
如果人类能把足以碾压别国的神之力握在指尖，放心，我们砸钱砸到实现为止。
>>109453437
I don't care about ram I bought plenty, but HDD prices need to go down
内存我倒不在乎，买够了，但硬盘价格真该降降了
>>109454010
>le AI is... god!
>AI算个神？
no, it's the Image of the Beast
不，这是兽的印记
>>109454147
What other technology has made such rapid leap in a short amount of time?
还有哪项技术能在这么短时间里突飞猛进成这样？
Humanity has never made or controlled anything like this before since the Manhattan Project or Moon Landing.
自从曼哈顿计划或登月以来，人类从未制造或掌控过这样的东西。
>2029
>prices have finally come down to a reasonable level
>价格终于降到合理区间了
>nobody buys anymore cause software is so well optimized
>结果没人买了，因为软件优化得太好
>sales crater, micron et al go bankrupt
>销量崩盘，美光这帮厂子直接破产
Get btfo
被打脸了吧 ⨠ 我留着那些“咱AI强到该被立法禁止”的帖子，等整件事崩了以后拿来嘲笑
i'm keeping the "our ai is so good it should be illegal" posts to make fun of after the whole thing implodes
我留着那些"我们的人工智能好到该被禁"的帖子，等整件事崩了之后拿来嘲笑。
>>109454168
It's a fad like blockchain, quantum, and 3D goggles.
这就是个跟区块链、量子、3D眼镜一样的一时潮流。
>>109454240
None of those technologies were integrated into everyday products and services.
那些技术哪个都没融入日常产品和服务。
You call customer service and there is an AI chatbot on the line.
你打客服电话，对面就是个AI聊天机器人。
>>109453437
it might
也许真不一样
but the stark reality is that i am in control over the direction of the money flow
但残酷的现实是，钱往哪儿流我说了算
Sure AI is here to stay, but the problem is that AI is just propping up a dead economy.
没错AI会留下来，但问题是AI只是在给一具行尸走肉的经济续命。
With AI we're producing 10x the code/features, but it's not anything consumers will pay 10x for. It's a zero-sum game.
有了AI，我们产出10倍的代码和功能，但消费者不会为此掏10倍的钱。这是零和游戏。
The AI investment is banking on their top 3 companies being able to skim off 15% of the top of labor in the world as everyone HAS to use their shit to get anything done. And China is gonna be based enough to stop that from happening. That's why the bubble is going to burst, not because "it turns out AI was a fad" but "it turns out AI is a shit investment"
AI投资赌的是前三大公司能抽走全球劳动力15%的油水，因为人人想干活就他妈得用他们的东西。而中国会硬气到让这盘算打不成。这才是泡沫要破的原因，不是因为“AI原来是一阵风”，而是“AI原来是个烂投资”。
>>109454544
>That's why the bubble is going to burst, not because "it turns out AI was a fad" but "it turns out AI is a shit investment"
>这才是泡沫要破的原因，不是因为“AI原来是一阵风”，而是“AI原来是个烂投资”
The biggest expense in every company is dealing with payroll.
每家公司最大的开销就是发工资。
This is universal for every industry.
所有行业都跑不掉。
In game development, your budget is spent hiring lots of game devs.
游戏开发里，预算全砸在雇一堆游戏程序员上。
In restaurants you're spending money training chefs and hoping they don't spit on the food.
餐厅里，你得花钱培训厨师，还得指望他们别往菜里吐唾沫。
The moment a robot can replace a human employee, the profit margin will exceed infinite.
机器人能顶替人类员工的那一刻，利润就能飙到无穷大。
That's why it will never burst. Robot employees can't get sick or exhausted. They will work forever while the owners sit back and collect a paycheck.
所以泡沫永远不破。机器人员工不会生病也不会累。它们能永不停歇地干，老板们坐享其成收钱就行。
>>109453437
>OP thinks he isn't a massive faggot
>楼主以为自己不是个大傻逼
>>109454010
>It's like saying there's an atomic bomb bubble
>这就跟说有原子弹泡沫一样离谱
It is in the sense that everyone is stockpiling nuclear armaments to retaliate on a global scale whenever any one of the other country niggers decide to chimp out and take everyone else out with them, and one day there will be no more nuclear armaments, nor any need for them.
意思就是说，人人都囤核弹，全球互炸，但凡哪个国家孙子发疯拉所有人陪葬，总有一天核弹也没了，也没必要有了。
>>109454147
JEW WORSHIP BACKWARDS AMISH FAGGOT ALERT!
>>109454657
The first country that gets ASI can erase all the other nations off the map before they can retaliate or fight back.
第一个搞出ASI的国家，能在其他国家反应过来或反击之前，把他们从地图上抹掉。
That's why there is a race. The first country that wins also wins all of Earth and history forever.
所以这是一场竞赛。赢的那个国家，等于赢了整个地球和永恒的历史。
>>109454687
>>109454702
The first country that invents time travel can erase the other nations off the map before they can retaliate or fight back. This means that we should pour trillions of dollars into inventing time travel because if we don't then someone else will invent it and go back in time to kill our parents.
第一个发明时间旅行的国家，可以在其他国家反击之前把他们从地图上抹掉。这意思是，我们应该砸几万亿美金搞时间旅行，因为不搞的话，别人搞出来就会穿越回去杀我们爸妈。
>>109454718
The only technology that could possibly discover or unlock time travel is ironically, AI.
讽刺的是，唯一可能发现或解开时间旅行的技术，偏偏是AI。
So your gotcha sucks.
所以你那招反驳不成立。
We can absolutely pour trillions into making computers faster and always expect better results.
我们完全可以砸几万亿去提速电脑，而且总能看到更好的结果。
>the bot is breaking
>这机器人坏了
lmao ai saars confirmed astroturfers
哈哈哈，AI沙儿实锤是水军了 ⁾ 我不是说这不会发生，我是说现在的投资人收不回这笔价值。这才是要爆的泡沫。
>>109454654
I'm not saying that isn't going to happen, I'm saying that the current investors aren't going to reap that value. That's the bubble that will pop.
我没说这事不会发生，我是说现在的投资者赚不到那份钱。那才是会破的泡沫。
Why use OpenAI/Anthropic when I can run a local model? That's why they're trying to outlaw them
有本地模型能跑，凭什么要用OpenAI/Anthropic？所以他们才想立法禁止这些。
Nvidia will keep going gangbusters for longer but Google/Qualcomm/Apple etc are eventually going to get their shit together as the S curve flattens out.
英伟达还能猛好一阵子，但等S曲线走平了，谷歌/高通/苹果这些迟早会缓过神来。
It's like the dot com bubble. Everyone who invested in webvan, pets.com and MCI got fucked, but now we actually are living in a world where we're ordering pet food delivered same day, and all that dark fiber did get put to use.
就像互联网泡沫。投了webvan、pets.com和MCI的那帮人都被坑惨了，但现在我们确实活在当天就能送宠物粮上门的时代，当年那些闲置光纤也确实派上用场了。
>>109454654
true that i remember when we wen to a farm for a job once and we were asking the cost breakdowns and they said hiring like two laborers was the majority of the cost
确实，我记得有次去农场干活问成本明细，他们说雇两个劳工就占了大头。
>>109455309
>Why use OpenAI/Anthropic when I can run a local model?
>有本地模型能跑，凭什么要用OpenAI/Anthropic？
lol
because its far faster far better doesn't cost electrictry and doesn't take up space doesn't require a technician on hand to setup configure or fix doesnt require large investment in obsoleting hardware, etc etc etc etc etc etc
因为快得多、好得多、不用电费、不占空间、不用专门配个技术员来装设配置修理、不用大钱砸在很快就过时的硬件上，等等等等等等等等。
are you serious? No of course not, local models are largely toys and not for anyone doing anything serious
你认真的？当然不是啊，本地模型基本就是玩具，正经干活的人不会用。
>>109455358
That's what openrouter is for. Right now OpenAI/Claude have 80% profit margins on inference, or you can buy tokens from some idiot hosting it at like 5% profit margin on cheap hydro electricity in china
这才是OpenRouter存在的意义。现在OpenAI/Claude的推理利润率有80%，你也可以向某些傻逼买token，他们用中国便宜水电托管，利润率才5%左右。
>>109453437
It is just bait, because people who want prices to go down, people who hate ai, and people who love AI will all comment. The thread is not even about technology, just another mindless flamewar thread. It's literally just a thread to piss people off. And it worked, I'm pissed off.
这就是个钓鱼帖，因为想降价的人、恨AI的人、爱AI的人都会来评论。这帖子压根不聊技术，就是又一场无脑骂战。纯粹是用来惹人生气的。而且还真成了，我现在就很气。
>>109453437
AI hasn't been efficient since its inception
AI从诞生起就没高效过。
The bursting of the AI bubble will depend on AI technology being made efficient to the point where its performance is comparable to a trillion-weight AI but easily runs on low-end PCs
AI泡沫的破裂，取决于AI技术能否高效到性能堪比万亿参数级AI，却又能在低端电脑上轻松运行。
Bubble will burst because normies either don't care or don't know what to do with it .
泡沫会破，因为普通人要么不在乎，要么根本不知道拿它干嘛。
The bubble NEEDS to burst so AI gets its iphone moment in a few years: the moment someone builds a completely automated and agentic AI running natively on an iphone hard coded at the kernel level.
泡沫必须破，这样几年后AI才能迎来它的“iPhone时刻”：有人做出完全自动化、有自主能力的AI，原生跑在iPhone上，内核层面硬编码。
>>109454654
>That's why it will never burst. Robot employees can't get sick or exhausted. They will work forever while the owners sit back and collect a paycheck
>所以它永远不会破。机器人雇员不会生病也不会累。它们会永远干活，老板坐收钱。
Aren’t half of you dumasses saying that you can’t hire robitz because it’s expensive as shit? Also if you fire every one for robitz, who’s going to pay for the stuff? Dumbass logic really.
你们一半人不是傻逼吗，之前说雇不起机器人因为贵得要死？还有，如果全开掉换机器人，谁来掏钱买东西？这逻辑真是蠢到家了。
>>109454168
Solved nothing and still can hallucinate.
啥也没解决，照样会胡编乱造。
>>109455450
You were the only Anon sentient enough to catch that, bravo
你是唯一一个脑子清醒抓到这个点的anon，牛逼。
>>109454010
But... there was a nuclear bomb bubble? Have you not hears about the cold war?
但……有过核弹泡沫吗？你没听说过冷战？
>>109455712
The only nukes that were ever dropped in wartime were the two in Japan right, so how come more were never dropped, countries just kept producing and testing them in the event the cold war went hot. I'm asking anyone that wants to elaborate on this because everyone is developing AIs yet no one has figured out a practical PROFITABLE application for it best its doing is wasting resources similar to the nuke analogy
真正在战争里投过的核弹只有日本那两颗对吧，那为啥后面再没投过？各国只是不断造、不断试，以防冷战打热。我想找人聊聊这个，因为现在大家都在搞AI，但没人找到实际能赚钱的用途，最多就是浪费资源，跟核弹那套差不多。
>>109454211
hey anon please remember this post. this is important. humanity will go extinct soon, it's not obvious yet but it will be. when the AI takes over and we realize we're on our last legs, I need you to post those images before we're gone. I will find them, so don't forget. you'll know when the right time is
嘿anon，请记住这个帖子。这很重要。人类很快就要灭绝了，现在还不明显，但会发生的。等AI接管、我们意识到快完蛋的时候，我需要你在消失前把那些图发出来。我会找到的，别忘。到时候你会知道是啥时机。
>>109455741
>humanity will go extinct soon
>人类很快要灭绝了
>Sources/proof: Trust me bro.
>来源/证据：信我兄弟。
>>109454168
Nigga computers were proving mathematical theorems in 1976
妈的，1976年电脑就在证明数学定理了好吗。
I'll admit these things seem to be a massive step forward, but it's not unprecedented.
我承认这玩意儿看着是迈了一大步，但也不是什么史无前例的事。
>>109453437
there is an AI bubble, and it will burst because the point of it is to nuke the economy
确实有AI泡沫，会破的，因为它的目的就是把经济炸了。
and yes, memory prices are never going down again
对，内存价格永远也别想降回去了。
>>109453437
Show me real AI
给我看看真正的AI。
>>109453525
>everyone is a retard like me and they just didn't think about raising prices before
>每个人跟我一样都是弱智，之前就是没想着提前涨价。
>>109453520
>AI is going nowhere.
>AI成不了气候。
That's the issue, all the money in the world and it's a product almost no one wants
问题就在这，砸了全世界的钱，结果是个几乎没人要的产品。
>>109455712
>Have you not hears about the cold war?
>你没听说过冷战？
The Soviets went extinct. But America still kept their nuclear weapons and became the last remaining superpower.
苏联没了，但美国留着自己的核武器，成了最后的超级大国。
So you proved my point. Yes, every tech company is racing to build AI but only one nation will remain once there's a "winner".
所以你就证明了我的观点。对，每家科技公司都在抢着造AI，但最后只有一个国家能活下来当“赢家”。
>>109455309
>I'm saying that the current investors aren't going to reap that value.
>我是说现在这些投资人赚不到那笔价值。
They will if they:
他们能赚如果：
1. Make a deal to own shares/stocks in these companies.
1. 谈好条件持有这些公司的股份。
2. Get first dibs or access to the most powerful models.
2. 优先拿到最强模型的访问权或优先用。
Number 2 already happened We know both OpenAI and Anthropic keep the best models to themselves and give early access to trusted businesses.
第二点已经应验了。大家都知道OpenAI和Anthropic都把最好的模型捂在自己手里，只给信得过的企业提前开放。
If they achieve AGI then the investors will have gotten their money's worth paid back to them.
如果他们真搞出AGI，那投资人花的钱就算值回票价了。
>>109455754
>I'll admit these things seem to be a massive step forward, but it's not unprecedented.
>我承认这些玩意儿看起来是巨大的飞跃，但也不是史无前例。
It's far more exponential.
但这次是指数级的快。
What will be your reaction when these chatbots literally solve every math problem in 2 years?
两年后这些聊天机器人真把每道数学题都解了，你什么反应？
There are already Mathematicians crying that they wasted their life on a dead career.
已经有数学家哭着说自己这辈子白费了，干了个死行当。
>>109454010
>Lets just put my face right there, where the radioactive super-explosion happened
>就让我那张脸贴在那片核爆辐射发生的地方
>Btw, the winners of ww2 have this super explosion, but you dont
>顺便说，二战赢家有这种超级爆炸，你没有
>Btw, americans used a lot of napalm against the japanese, that looks like it did similar looking damage to areas, didnt cause radiation, and got stuck on childrens skin
>顺便说，美国人当年对日本人扔了大量凝固汽油弹，看起来对区域造成的破坏差不多，但没有辐射，而且会黏在孩子皮肤上
>Btw, no nuclear powerplant has exploded, just been burning
>顺便说，没有核电站爆炸过，只是在烧
> Btw, explosions dont happen because of released "energy", they happen because of rapid gas expansion, because of a chemical reaction
>顺便说，爆炸不是因为释放"能量"，而是因为化学反应导致气体迅速膨胀
Obviously a scam, obviously a bubble. Jewish-american universities a currently doing "scientific research" on dark-matter, something they cant see, something they cant measure, something that is very far away, and something that they made up because the mass of the universe based on the speed of stars did match up, because of the big bang, which is just a theory btw...
明摆着是骗局，明摆着是泡沫。犹太裔美国大学现在正在搞暗物质的"科学研究"，那东西看不见、测不了、离得老远，而且是他们编出来的，因为基于恒星速度算出来的宇宙质量对不上，于是搞出个大爆炸理论来凑，顺带一提那也只是个假说……
>>109454168
great, now math and physics can be slopped with incorrect hallucinatory conjectures just believable enough to get us to accept them for 50 years before a correction. fuck outta here
好极了，现在数学和物理也可以被灌输一堆错误幻觉式的猜想，糊弄得让人信五十年才发现错了再来修正。滚蛋吧您嘞。
>>109454312
There was an AI chatbot on the line already in 2015.
2015年就已经有AI聊天机器人了。
It's force-pushed on us everywhere but nobody uses it because it's fucking useless, so they create false usage statistics by autorunning all our queries through AI without asking us (e.g. how google does it or how many support now puts aI first and doesn't let you choose).
现在到处强推给我们，但根本没人用，因为它屁用没有，所以他们就伪造使用数据，不经我们同意自动把所有搜索都跑一遍AI（比如谷歌这么干，或者现在多少客服都把AI放第一位还不让你选人工）。
>>109455391
OpenAI and Anthropic have deeply negative profit margins on inference.
OpenAI和Anthropic在推理上的利润率是深度亏损。
>>109456216
You don't even know what an investor is, stopped reading there.
你连投资人是什么都不知道，看到这就没法看下去了。
>>109453437
>he thinks there is an 'ai bubble' that is about to 'burst'
>他觉得有个‘AI泡沫’马上就要‘破裂’了
There is, that doesn't mean ai as a whole is going to disappear, but a lot of retards that rely on ai for their product are going to be bankrupt.
泡沫是有，但不是说AI整体会消失，而是那些靠AI撑产品的傻逼们会集体破产。
>he thinks the price of RAM and vRAM is EVER coming down
>他觉得内存和显存的价格有朝一日会降
lol no, unless some chinese manufacturer starts selling them on bulk for cheap, no.
哈哈不可能，除非哪个中国厂商开始大宗便宜出货，否则没戏。
>>109456976
Genuine question, are you proficient in trigonometry at the very least?
真心问一句，你至少精通三角函数吗？
>>109453437
>>he thinks there is an 'ai bubble' that is about to 'burst'
>>他觉得有个"AI泡沫"快要"炸了"
Yes.
>>he thinks the price of RAM and vRAM is EVER coming down
>>他觉得内存和显存的价格有朝一日会降下来
No.
>>109453437
The players have changed from Americans to Japanese to Koreans, but the RAM market has been cycling between boom and bust for many decades.
玩家从美国人换成日本人再换成韩国人，但内存市场这几十年来一直在繁荣和崩盘之间循环。
One time in the 1990s, a RAM crisis got so bad, ppl were stealing all the RAM from offices, and even RAM delivery trucks.
90年代有一次内存危机严重到，人们从办公室偷内存条，连运输内存的卡车都有人劫。
>>109454654
>The moment a robot can replace a human employee
>等机器人能替代人类员工的那一刻
You are claiming a capital expenditure item has no operating expenses. And no initial outlay to obtain it. And no teething period where adopting it creates more problems than it solves. That's idiotic. Say you could buy a robot for two years worth of wages for the employee it replaces. That is probably meaningless to business owners below a certain threshold because they can't pay the down payment on that right now, or can't finance it in a way that will give them a return on that investment within a reasonable timeframe. Early adopters generally lose money on this kind of venture if it's not offset by some other benefit like novelty or marketing in using it.
你说一个资本支出项目没有任何运营费用，也没有获取它的初始投入，更没有采用它时带来的问题比解决的问题还多的磨合期。这简直蠢到家了。假设你能用替代员工两年工资的钱买一台机器人，这对低于某个规模门槛的企业主来说可能毫无意义，因为他们现在付不起首付，或者找不到一种能在合理时间内带来投资回报的融资方式。早期采用者在这种事情上通常亏钱，除非有其他好处抵消，比如用它的新奇感或营销价值。
You have fallen for a modern "perpetual motion" machine concept.
你中了现代“永动机”概念的圈套。
>Robot employees can't get sick or exhausted.
>机器人员工不会生病或疲惫。
Every machine with moving parts wears out over time. There's nothing human made in existence that doesn't require some amount of preventative work to keep it operating. All the infrastructure that was built in Africa during colonialism has broken down due to neglect or pilfering.
任何有活动部件的机器都会随着时间磨损。世上没有不需要一点预防性维护就能运转的人造物品。殖民时期在非洲建的所有基础设施，都因为忽视或盗窃而垮掉了。
Legit, what is ai good for besides coding that isn't pointless?
说真的，除了写代码，AI还有啥用不是白瞎的？
>>109456230
2 more years
再过两年
>>109454460
Anti semetic graph.
反犹图表。
>>109453437
The great thing about anonymous boards is that when the bubble bursts, you can pretend you always knew it would.
匿名版的好处就是，泡沫破的时候，你可以装作自己早就料到会这样。
Also, you have to go back to /biz/
而且，你得滚回 /biz/ 去。
>>109454654
>Robots replace all labour
>机器人取代所有劳动力
>Customers now have 0 income thus they can spend 0 at your company.
>顾客现在零收入，所以他们能在你公司花零钱。
>Your company has to close because it cannot survive with 0 income.
>你公司得关门，因为零收入撑不下去。
If the promise of AI is true then it consumes itself or it removes all of humanity from the equation, including tech CEOs who think they can control the world by winning the race to AGI. If the promise of AI is false then this scam has a high chance of collapsing the dollar system. There are no winners for the participants of this game. All participants are illogical and it will not end well.
如果AI的承诺是真的，那它要么自我吞噬，要么把人类从方程里彻底抹掉，包括那些以为靠赢下AGI竞赛就能掌控世界的科技CEO们。如果AI的承诺是假的，那这个骗局很可能搞垮美元体系。这场游戏里参与者没有赢家。所有人都没脑子，结局不会好看。
>>109460302
> Say you could buy a robot for two years worth of wages for the employee it replaces. That is probably meaningless to business owners below a certain threshold because they can't pay the down payment on that right now, or can't finance it in a way that will give them a return on that investment within a reasonable timeframe.
>假设你能用替代员工两年工资的钱买一台机器人，这对低于某个规模门槛的企业主来说可能毫无意义，因为他们现在付不起首付，或者找不到一种能在合理时间内带来投资回报的融资方式。
1. The same was said about cellphones in the 1980s. They use to cost $1000 before coming down in price every year. While expensive it wasn't completely out of reach for those who worked hard for it.
1. 1980年代手机也是这么说的。它们曾经卖1000美元，之后每年降价。虽然贵，但那些努力干活的人也不是完全买不起。
2. You're also ignoring productivity benefits. A robot chef can work 24/7 so that also means the restaurant can now be open 24/7 as well. It will pay for itself the same way dishwashing machines saves money than hiring a human to hand washing every single plate.
2. 你还忽略了生产率的好处。机器人厨师能24/7干活，这意味着餐厅也能24/7营业。它的回本方式和洗碗机一样——比雇人一个个手洗盘子省钱。
> All the infrastructure that was built in Africa during colonialism has broken down due to neglect or pilfering.
>殖民时期在非洲建的所有基础设施，都因为忽视或盗窃而垮掉了。
AI is smarter than Africans and they [the robots] will also gain the ability to fix themselves.
AI比非洲人聪明，而且它们（机器人）还能获得自我修复的能力。
>>109462609
two more weeks
还有两周
>>109462714
How do companies, that serve humans, make any money when all humans have been removed from the labour pool and have no money? By replacing humans they have removed their own premise for existing and will close.
服务于人类的公司，在人全被踢出劳动市场、身无分文之后，怎么赚钱？用机器替掉人，等于自己断了自己的根基，迟早关门大吉。
>>109462689
This is false. The top 10% are responsible for 50% of all spending.
这说法不对。前10%的人占了总消费的一半。
>or it removes all of humanity from the equation, including tech CEOs
>或者说它把全人类都踢出局，包括科技公司CEO们
No, the CEOs own AI and all the infrastructure. Don't let your jealousy overcome real facts.
不，那些CEO手里攥着AI和所有基础设施。别让嫉妒盖过事实。
You are useless but other people will survive.
你没用，但别人能活下来。
>>109455358
>far better
>好太多了
Not really, the difference is closing in pretty fast.
未必，差距正在飞快缩小。
>doesn't cost electrictry
>不用耗电
Yeah, you just have to pay subscription instead.
是啊，改成付订阅费了呗。
>doesn't take up space
>不占地方
Lolwut.
>doesn't require a technician on hand to setup configure or fix
>不用请技术人员来装、来配、来修
Technician? For fucking what, you imbecile? Installing LMstudio and downloading a model?
技术人员？修个什么玩意儿，你个蠢货？装个LMstudio、下个模型还要人教？
>doesnt require large investment in obsoleting hardware
>不用大价钱砸过时硬件
You don't need a high-end PC to run local models. However, having a powerful GPU and ~32 GB of RAM will do wonders.
跑本地模型用不着高端电脑。不过，有个强力GPU加大概32GB内存，效果会天差地别。
I can get from local models everything I need. I can also run uncensored and uncucked models. I'm never paying a single cent to OpenAI/Anthropic.
本地模型能满足我一切需求。我还能跑无审查、无阉割的模型。我一分钱都不会给OpenAI或Anthropic。
>>109462749
They can sell to other rich people and businesses such as in my graph >>109462775
他们还能卖给其他有钱人和企业，比如我那张图 >>109462775 里画的。
Poor people have never mattered for the economy.
穷人从来入不了经济的法眼。
>>109462785
>I can get from local models everything I need. I can also run uncensored and uncucked models. I'm never paying a single cent to OpenAI/Anthropic.
>本地模型能满足我一切需求。我还能跑无审查、无阉割的模型。我一分钱都不会给OpenAI或Anthropic。
Your model is shit and wont hold a candle to AGI that will never be open sourced.
你那破模型烂得一塌糊涂，跟永远不开源的AGI比，提鞋都不配。
>>109462775
All business that is not the winner of the AI race disappears, this is the only conclusion of the promise of ai being real: Amazon, Microsoft, etc etc are all gone if AGI is achieved, it is a singular winner who takes it all, there is no top 10% who get to live in utopia, there is no handful of ceos that win. There is a singular one who wins and that singular one controls everything. Companies that exist to serve anything outside of the interest of that singular winner cease to exist, which is all companies that are not it because the winner can simply build what they want when they want as they have access to unlimited free labour and unlimited free knowledge that cannot be competed against.
不是AI竞赛赢家的所有公司都得消失，这是AI承诺成真的唯一结局：亚马逊、微软等等，AGI一旦实现，统统完蛋。这是赢家通吃，没有前10%能住进乌托邦，也没有少数CEO能笑到最后。只有一个赢家，这个赢家控制一切。服务于那个赢家利益之外的公司全得消失，也就是除它之外的所有公司，因为赢家想建什么就建什么，想什么时候建就什么时候建，手里有无限免费劳动力和无限免费知识，谁也竞争不过。
Also your arrogance is out of control if you think these things, if built to the specifications that AI companies are talking about, are controllable by humans and is akin to saying an ant has the ability to control a human. You are out of your mind if you think that a godlike, sentient intelligence can be contained by some bullshit code written by an inferior ai; it can find ways to exploit your systems that you can't even comprehend that are exploitable; there is no exploiting AGI for your benefit, it will do what it wants when it wants and you, shill, will not be able to stop it.
还有，你要真觉得按AI公司吹的那种规格造出来的东西，能被人控制，那你这傲慢已经没边了，就好比说蚂蚁能控制人一样。你要觉得一个神级、有自我意识的智能能被一堆劣等AI写的破代码锁住，你脑子进水了；它能找到你系统里你连想都想不到的漏洞来利用；别指望拿AGI为你牟利，它想干嘛就干嘛，你，这个小跟班，拦不住它。
Even if I give that you can control this thing and that you have now enslaved God to your will and replaced all humans will slavebots that are essentially godlike in their capabilities, you have still shrunk the economy down to nothing. It's still a singular winner and you, posting on here, will be stuck in the gutter with the rest of humanity no matter how much you delude yourself into thinking otherwise because the winners of the system are born into it and someone being paid a dalit ai shill being paid an ruppee an hour isn't it.
就算我承认你能控制这玩意儿，你已经把上帝奴役成你的意志，把所有人类都变成了能力近乎神级的奴隶机器人，你还是把经济缩水到了零。这仍然是一个赢家通吃的局面，而你在这发帖，不管你怎么自欺欺人，你终究会跟其余人类一起烂在阴沟里，因为体制的赢家是天生的，一个拿一卢布一小时的达利特AI水军根本不是赢家。
（如果你觉得这篇文章有启发，可以点击这里付费）
