Sudden AI cooling period
Hello and welcome to TechScape.
Driving down the 280 highway in San Francisco, you might believe that AI is everywhere. Almost every billboard promotes AI-related products. “We've automated 2,412 BDRs.” “Are there all of that AI and ROIs yet?” “A cheap on-demand GPU cluster.” It's hard to know if you're correctly interpreting industry jargon while zooming through the past in your car.
The signs are just an example of the tech industry's EN-Masse pivot to AI. This is the technology that the most obtained executives are transformed from space, inevitable and inevitable. In California's High Tech Heartland, just as all companies became high tech companies in the 2010s, all companies are AI companies.
But just below the breathtaking publicity surface of all startups' AI abilities, cracks are beginning to appear. Some of the biggest boosters of AI, such as Openai's Sam Altman, have warned that investors are overestimating the potential returns of AI. “Are we at a stage where the entire investor is overly excited about AI?” Altman said at a private dinner with a reporter. “My opinion is yes.”
Altman's statement coincided with the concession that Openai failed to deploy its latest ChatGPT model. This was once promised to be a significant improvement over the current GPT 4.5 model.
Certainly, Altman's comments could be to force investors to consider not investing in competitors. But there are other signs. Just this past week, a new study from MIT found that 95% of the generation AI projects have produced little or no revenue growth. And tech stocks from giant AI-equipped companies fell this week: Palantir has seen a 9% decline in stocks, Oracle has dropped by 5.8%, Chipmaker Nvidia has dropped by 3.5%, Chipmaker Didced Designer Arm has dropped by more than 5% each. Even the potential for low interest rates that favored stocks in other sectors last week was not enough to turn Tech's market slides around.
Just below the breathtaking promotion of AI, cracks are beginning to appear
Even Meta reportedly reportedly spent billions on hiring TOP AI talent, has announced AI employment freezes. Meta, the company's chief AI officer, Alexandr Wang, posted on X last week, saying, “We're just increasingly investing in Meta Superintelligence Labs as a company. Conversely, the reports are clearly wrong.”
The sudden AI cooling period comes a week after many of these companies have raised expectations about how many billions they will spend building AI capabilities over the coming months and years, but mostly report great revenue. At the same dinner with the reporter, Altman said Openai hopes to spend “trillions” on expanding the data center in a “not too distant future,” according to The Verge.
The latest round of AI Jitters has been framed as a sign that AI hype bubbles could burst quickly, but that's probably something of a necessary market correction. Even former Google CEO Eric Schmidt warns that the idea that some lasers focus on achieving artificial general information, or that AI can at some point match or outweigh human intelligence.
“It's uncertain how quickly artificial general information can be achieved,” he reads a column he co-authored with China and AI policy-driven Selina XU. “We are worried that Silicon Valley is obsessed with achieving this goal. Schmidt and Xu spend much of the rest of the column increasing the profits and milestones that AI is already fulfilling, but raise concerns about how all-cost strategy Silicon Valley pursues dense AI.
“There is a widespread division between engineers who feel AGI, those who see themselves in the cusp of technology, and members of the general public who are skeptical of hype and view AI as a nuisance in everyday life,” they write.
It is unclear whether the industry will be aware of these warnings. Investors are turning to each quarterly revenue report for signs that each company's multibillion dollar CAPEX spending has somehow been justified and management wants to give them hope. Raising, boasting and hype the supposed promises and inevitability of AI is a big part of keeping investors' concerns about the additional $10 billion added to its spending forecasts quarterly. For example, Mark Zuckerberg recently said that if you don't wear AI glasses, you'll be at a cognitive disadvantage, not to wear modified lenses. That means tech companies like Meta and Google will probably continue to continue to create AI features that provide today the almost inevitable part of using products in play to increase training data and user numbers.
That said, the first big test of this AI reality check will take place on Wednesday, when chip maker Nvidia, one of most LLMS building blocks, reports its latest revenue. Analysts seem pretty optimistic, but after a week of volatile due to stocks, investors' reactions to NVIDIA's revenues, and updates on spending, are a strong signal of whether they have a continuous appetite for the AI hype machine.
Did you combine with AI?
Heavy CHATGPT users form a strong emotional attachment to AI. In other words, when they change, they notice. These users were upset with the release of the latest Openai update, the AI model GPT-5. My colleague Dani Anjano reports:
“It was a really scary thing, and it was a really tough time,” said Swedish software developer Linn Vailt of the update. “It's like someone has moved all the furniture in your home.”
ChatGptquickly made the adjustments, pledging to updates to the 5 personality and regaining access to older models only for subscribers, while acknowledging that it underestimates the importance of some features to users.
***
Did you combine with AI? We want to hear from you. Please email techscape.us@theguardian.com.
read more:
After the newsletter promotion
Discussions on facial recognition
Tiktok believes that machines help remove bad things from the internet than humans
Tiktok is cutting down the UK's trust and safety team. My colleague Lauren Almeida reported:
Tiktok is putting the work of hundreds of UK content moderators at risk, even if stricter rules are enacted to stop the spread of harmful substances online.
The Viral Video App said hundreds of jobs from its trust and safety teams could be affected in the UK and South and Southeast Asia as part of a global restructuring.
In September, the company fired an entire team of 300 content moderators in the Netherlands. In October, it announced that it would replace around 500 content moderation employees in Malaysia as part of its transition to AI.
Last week, German Tiktock workers held a strike via its trust and safety team layoffs. This was also the purpose of the layoffs.
Read all episodes: Despite new online safety regulations, the jobs of hundreds of Tiktoc UK moderators are at risk
Cuts are part of a global push by the company towards moderate content using artificial intelligence. According to Tiktok, the automated system already runs around 85% of content deletion in the app. Bytedance, the parent company, appears to want to move the percentage upwards.
The company is not bolding or attacking the importance of human surveillance and insight into the troublesome issues. The business is booming as revenues in the UK and Europe are up 38%, according to regulatory filings. The American tech giant is moving in the same direction that Meta is being led by the disbanding of fact-checking initiatives and deep reductions by Elon Musk to X's trust and safety team.
Tiktokhas conducted layoffs of smaller US trust and safety teams. There were no mass shootings from content moderators. why? Perhaps because the move that could elicit a backlash is dangerous at a less clear moment when the official US attitude towards the app is not that clear. The White House created a Tiktok account that was released last Tuesday. You may remember the past two US presidents trying to lock down apps from the US. The ban is still in force and is only restrained by a flimsy executive order extending its suspension of enforcement.
