00:00A reminder before watching the video
00:01There's more content on trending websites.
00:04Let me ask you a question first.
00:05If one day AI starts creating its own AI
00:09Are you scared?
00:10So this actually happened.
00:13It was in early June.
00:14American artificial intelligence company Anthropic
00:17Suddenly a major report was released.
00:19Warnings say that AI has now begun to exhibit some...
00:22Signs that it can escape human control
00:24Their latest AI model
00:26A disturbing trend has already emerged.
00:29AI is accelerating the creation of the next generation of even stronger AI.
00:34The report also warned
00:35Artificial intelligence systems are approaching a tipping point.
00:39They may automatically develop their own successors.
00:42In this report...
00:43What makes the industry most uneasy
00:45AI has the ability to optimize itself.
00:47Upgrade yourself
00:48No longer requiring much human intervention.
00:52They even have the ability to design
00:53The next generation that is more powerful than themselves
00:56Faced with this risk of spiraling out of control
00:58entropics rarely stand up
01:00Calls for cutting-edge AI labs to establish a coordinated system
01:04Verifiable mechanism
01:05Consider slowing down if necessary
01:08Even suspending the development of advanced AI systems
01:11What it wants is not for a single company to stop on its own.
01:14Instead, it is a globally coordinated mechanism.
01:16To achieve a meaningful pause
01:19It requires several well-resourced and cutting-edge laboratories to work together.
01:23agree
01:24There should be clear rules specifying what situations will trigger this.
01:27Under what circumstances can the pause be lifted?
01:29There also needs to be an independent body to oversee it.
01:32So what gives them the right to urge everyone to stop?
01:35Because the internal data is somewhat frightening.
01:38The company said that as of May 2026
01:41More than 80% of the code has been merged into their codebase.
01:45It was written by their model, Cloud.
01:49Before Cloud Code launches in early 2025
01:52This percentage is still only in the single digits.
01:55In other words, in just over a year
01:58They transformed AI from a menial task provider into a primary program writer.
02:02Simply talking about code ratios might not be enough to make a significant impact.
02:05So they simply conducted a test.
02:07Give the cloud a program to train a small AI.
02:10He was required to make no mistakes.
02:13Optimize the running speed as fast as possible.
02:16According to reports
02:17This question was posed by a skilled human researcher.
02:20It takes 4 to 8 hours
02:22It only managed to achieve about 4 times.
02:24And the final result is
02:26AI, however, surged to 52 times.
02:30This is not just a little bit faster.
02:32They are not even in the same league.
02:36But after hearing so much
02:38People might find it a little scary.
02:39But there is one thing
02:40This is a very interesting point to consider.
02:42Anthrobic just completed a $65 billion...
02:45Series H funding
02:47Valuation soars to over $900 billion
02:50Even surpassing OpenAI
02:52This industry leader
02:53This year's annualized revenue exceeded US$47 billion.
02:57Furthermore, they secretly filed a report with the U.S. Securities and Exchange Commission.
03:00Improved listing application
03:02At this time of preparing for listing
03:05Suddenly ran out and shouted
03:07Wait a minute
03:07Very dangerous
03:08Everyone, stop for a moment.
03:09Do you guys think there's something wrong with this logic?
03:12so what
03:13Many investors and market observers have doubts.
03:16This was a meticulously planned marketing stunt.
03:21Just thinking about it.
03:22Like the capital market claims
03:23Our AI has become so powerful that we must halt its development.
03:27What could be a better way to boast about the technological prowess of one's own products?
03:32It's no wonder that outsiders feel they are ostensibly sounding a warning.
03:36In reality, it's a publicity stunt for their upcoming IPO.
03:40Push your personal worth to its peak
03:43What about the related conspiracy theories?
03:45There is another very convincing point of view.
03:47It was in the same week that Anthropic published his article.
03:50His arch-rival OpenAI also happened to publish a paper.
03:53An article titled "Blueprint for AI-Driven Democratic Governance from Two Years Ago"
03:58Similarly, AI self-improvement is listed as a key area that must be controlled in the future.
04:02force
04:03What about the two leading companies?
04:05Almost simultaneously, they raised the exact same topic of panic.
04:09What are they planning?
04:12Informal advisor to US President Trump
04:14Venture capitalist Sachs fired back.
04:17The criticism of Enthropy's actions is arguing that they are essentially promoting regulatory capture.
04:22Simply put, these large companies are trying to create panic.
04:26To mislead the government into introducing stringent regulatory policies.
04:29This limits the development of low-cost AI models.
04:33This naturally consolidates its monopoly position.
04:37Faced with this situation where you can run fast
04:40But what about having someone else step on the brakes?
04:43What about Nvidia founder Jensen Huang?
04:45He once publicly joked that
04:47Some people develop a sense of divine intervention after becoming CEOs.
04:52He meant that they had begun to worry about the world.
04:57But he himself already possessed more weapons than others.
05:01Whether others dare to come up is another matter.
05:04Regardless, a professor at the University of Pennsylvania
05:07Professor Morrick has a different opinion.
05:09He believed it was indeed Enthropy's warning.
05:13There may be some marketing elements.
05:15However, the concerns about AI going out of control are also real.
05:19It is something that all of humanity should take seriously.
05:23This dramatic call for the world to apply the brakes
05:25Is it truly out of a genuine concern for the future of humanity?
05:29Or is it a carefully calculated business strategy by the giants?
05:32That's something for everyone to think about.
05:43Thank you for watching
Comments