Sam Altman’s Ai Paradox: Warning from a bubble while generating trillion

I was not invited to a comfortable Sam German dinner with correspondents in San Francisco last week, but this may be the best. I face a problem in suppressing angry eye rolls when I hear peak data in the silicon valley -.

I am not sure that I could control myself when the CEO of Openai said he believed that artificial intelligence could be in a “bubble”, with market conditions similar to the dotcom mutation in the 1990s. Yes, it is According to what he said, he said“Investors as a whole are exaggerated about artificial intelligence.”

However, on the same meal, Altman also said that he expects Openai trillion dollars to build on his data center in “The Future is very not far”, adding that “you must expect a handful of economists who wander in their hands, saying:” This is very crazy, it is very reckless, and we will only be like, “You know what?

UMM … What could be strict than a billion dollars expansion mode in an industry that you just described as a bubble? Eye roll braid reaches the top of my head. Altman may definitely refer to the smaller startups of artificial intelligence with high assessments in the sky and slim or non -existent revenues, but it is still irony to be rich. It is worth noting in particular, given the weak GPT-5 offer earlier this month, which was supposed to determine a jump forward, but instead left a lot of his disappointment Steering system And no progress.

In addition, even when Altman talks about bubbles, Openai itself raises record sums. In early August, Openaii received $ 8.3 billion from new financing, with a rate of $ 300 billion – part of its plan to raise $ 40 billion this year. This number was five times in the subscription. Moreover, employees are now preparing to sell about $ 6 billion of shares for investors such as SoftBank, Dragoneer and Thrive, which causes the company’s evaluation to $ 500 billion.

Openai is hardly strange in its infrastructure. Technology giants are pouring unprecedented amounts Amnesty International Invision In 2025: Microsoft It plans alone to spend $ 80 billion on artificial intelligence data centers in this fiscal year, while Dead It offers up to $ 72 billion in artificial intelligence and infrastructure investments. On the FAFA Front, Openai also has a company-competitors like Antarubor chasing their billions of dollars.

It seems that the largest bulls in Wall Street, like Dan Evez of Wedbush, are not interested. Evis said on Monday on the “closing bell” from CNBC that the demand for Amnesty International’s infrastructure has grown by 30 % to 40 % in recent months, describing CAPEX at the moment of health verification of the sector. While “some butter” admitted in parts of the market, he said that the artificial intelligence revolution with independent systems began only to play and we are in the “second half of the nine games match.”

While the bubble involves a final explosion, and all the damages that produce, the main phenomenon that causes the bubble often has real value. The appearance of the web was in the 1990s; The bubble was a reflection of the huge opportunities that opened.

However, I will be curious if anyone presses Altman on the paradox of artificial intelligence, which gets a bubble while simultaneously boasts of collecting massive donations in Openai. Maybe above a cup of champagne and sweet sugary candy? I also like to know if he has provided more striking questions about other big issues that wave on the horizon on the company: turning it into the public benefits company (and what this means for non -profit organizations), the current situation of its partnership in Microsoft is still, and whether its mission is “AGI to benefit from human”.

In any case, I am a follow -up chat game with Altman & Co (call me!). I will bring champagne, turn off the questions, and do my best to keep the eye rolls in a critical position.

Also: within just a few weeks, I will go to Park City, Utah to participate in our annual conference technology in Montage Deir! The space is limited, so if you are interested in joining me, Sign here. I highly recommend: There is a wonderful group of speakers, including Ashley Kramer, chief revenue in Openai; John Werner, President and CEO of Walmart Us; Tony Show, Founder and CEO of Doordash; And a lot, a lot!

However, here is more news of artificial intelligence.

Sharon Goldman
[email protected]
Sharongoldman

Our wealth

Wall Street is not concerned about the artificial intelligence bubble. Sam Taman Written by Beatrice Nolan

Massachusetts Institute of Technology Report: 95 % of the artificial intelligence pilots in companies fail Written by Sherrill Estrada

The Silicon Valley talent continues to recycle it, so this executive director uses the “MoneyBall” approach to reveal the geniuses of artificial intelligence hidden in the new era Lake Sydney

Waymo is an experience with the IXC, but EXEC says Lidar and Radar sensors are important for self -driving safety “in all circumstances” Written by Jessica Matthews

Amnesty International in News

More shocks for Meta Ai. the New York Times I mentioned Today Mita is expected to announce that she will divide the Amnesty International Department – known as Meta Super WABS LABS – into four groups. One will focus on artificial intelligence research; One on “Superintelligence“The last of the products; and one on the infrastructure such as data centers. According to the unknown sources, it is likely to reorganize“ the final for some time, ”with movements” aimed at better organizing Mark Zucker, and better developing AI products. Repaired Amnesty International, including the CEO of AI AI AI, was widely brought as head of artificial intelligence employees.

Madison Street began to love artificial intelligence. According to New York Times,Artificial intelligence calmly became a key player in advertisements. What I felt when Coca-Cola released a holiday advertisement created from artificial intelligence last year is now the prevailing: Nearly 90 % of large budget marketers are already using-or planning to use artificial intelligence in video ads. From excessive backgrounds to artificial audio operations, technology reduces costs and production times, and opens TV spots to smaller companies for the first time. Companies like Shutlerock and ITV help brands in replacing weeks of work with hours, while technology giants like Meta and Tiktok pay their AI’s advertising tools. This shift raises ethical questions about the displacement of creators and deceptions, but industrial leaders say that the genie outside the bottle: Amnesty International not only wear advertisements – it reshapes the entire commercial play book.

Silicon Valley AI’s deals create startups Zombie: “I have caught the institution.” according to CNBCand The start -up scene of artificial intelligence is cavity in Silicon Valley as rules for combating Big Tech with a new PlayBook book: Licensing deals and talent raids that young companies belong. Windsurf, once in conversations to be obtained by Openai, has collapsed into turmoil after its founders have passed to Google in a $ 2.4 billion license agreement; The interim CEO Jeff Wang described the meetings of all hand as tears because the employees realized that they had left “nothing.” Similar movements have seen Meta 14.3 billion dollars in Scale AI, Microsoft Scooop Up Forders’ Founders, and Amazon Strip Talent from ADEPT and Covariant- which is called the so-called “zombies” with a few future. While the founders and senior researchers scream, investors, ranks and files have been cut off, which are increasingly concerned that these semi -air progress are not only hiking, but also threatening to suffocate artificial intelligence in their source.

Sources say NVIDIA is working on the new AI chip for China, which surpasses the H20. according to Reutersand NVIDIA is developing China’s new Amnesty International chip, named B30a, based on the advanced Blackwele structure. The chip, which can be delivered to Chinese clients for the test, will be stronger than the current H20 but is still located below the American export sills-using one environment design with about half of the raw computing power of the pioneering NVIDIA B300. The move comes after President Trump pointed to a possible approval of the sales of the bizarre chips to China, although organizational approval is unconfirmed amid fears of the two parties in Washington for giving Beijing access to advanced AI devices. NVIDIA argues that keeping Chinese buyers is crucial to prevent the defection on local competitors such as Huawei, even when Chinese organizers make doubts about the company’s products.

Eye on artificial intelligence research

The study finds that the interviews led by artificial intelligence are improved. A New study Look at what is happening when job interviews are performed by artificial intelligence sound agents instead of human recruits. In a big experience with 70,000 applicants, people were randomly assigned to conduct interviews by a person, by artificial intelligence, or give the selection. Surprisingly, the interviews led by artificial intelligence have already improved the results: the applicants that artificial intelligence met 12 % more likely to get job offers, more likely to start jobs, and more likely to employees by 17 % after 30 days. Most applicants did not mind change – even 78 % chose artificial intelligence when giving the option, especially those who have fewer test scores. Artificial intelligence also brought out more useful information than the candidates, prompting employment companies to evaluate those interviews higher. In general, the study shows that the interviews of artificial intelligence can completely perform, or even better than human recruits – without harming the consent of the applicants.

You have a calendar

September 8-10: Fortune Brainstorm Tech, Park City, Utah. Apply to attendance here.

6-10 October: world Amnesty International The week, Amsterdam

October 21-22: Teddy San Francisco. Apply to attendance here.

2-7 December: NeuPIPs, San Diego

December 8-9: Fortune Brainstorm Ai San Francisso. Apply to attendance here.

Brain food

Should Ai Chatbots be protected from damage?

Artificial Intelligence Anthropier Laboratory A new safety scale In the latest Claude models, which enable artificial intelligence to end the conversations in extreme cases of harmful or abusive interaction. The advantage is only activated after the frequent re-direction-badly for content requests that involve sexual exploitation of minors or to facilitate violence on a large scale. The company is significantly working to frame this as not mainly protection for users, but for the “artificial intelligence luxury” of the model, which reflects an exploratory position on the potential moral situation of the device.

It is not surprising that the idea of giving an ethical state of Amnesty International is controversial. Jonathan Bersh, a professor of philosophy at London Economy College, said, Guardian He welcomed the Antarbur step to stir a public debate about the famous intelligence – the topic that many in this industry prefer its suppression. At the same time, he warned that the decision risks misleading users in the belief that Chatbot is more realistic than it is.

Others argue that focusing on the well -being of artificial intelligence distracts from urgent human concerns. For example, although Claude is designed to finish the most abused conversations, he will not interfere in imminent self-harm-though a New York Times Yesterday urged such guarantees, which she wrote, or did she discover Chatgpt talks only after her daughter committed suicide.