If AI doomerism is freaking you out, consider the advice of Yann LeCun, former Meta AI chief and a scientist with over 40 years of experience in the field.

Why it matters: As CEOs warn of job loss and existential risk, Turing Award-winning LeCun argues the real danger is making life-altering decisions based on exaggerated claims about the future of the technology.

The big picture: The leading AI scientist warns that doom narratives are already harming teens' mental health, calling extinction fears "extremely destructive" — and wrong.

• "A small proportion of high school students are actually kind of depressed because they've read that AI is not only going to take a job, but basically cause human extinction," he said. "They take that seriously and it has a profound effect on their psychology."

No, AI won't eliminate 20% of jobs, and yes, you still need to go to college, LeCun argues. Here's more advice from LeCun, executive chairman of AMI Labs:

1. Don't listen to CEOs

• It can be hard to keep up with the AI race when CEOs of the frontier labs keep saying each model is more powerful and potentially world-ending than the last. LeCun tells Axios: Ignore them.
• "Don't listen to CEOs ... they have a vested interest in propping up the power of the products they sell," he says, adding that AI CEOs are also not the ones to listen to about the impact of AI on labor. That's a job for economists.
• AI tools are powerful, but "still not very good at reasoning," and "there is a long history ... of researchers in AI having a widely optimistic view of when machines will become more intelligent than humans ... this is not going to take us to human-level AI for quite a while."
• LeCun's view is reflected in his work as executive chairman of AMI Labs, where he's building AI systems aimed at overcoming the reasoning limits he sees in today's models.

2. Go to college

• Sorry, kids. LeCun has spoken, and says you still need to go to college. He argues there's more value in advanced degrees because AI will increase demand for more educated, critical thinkers.
• "Study things with a long shelf life," he said, which is, of course, hard to parse. He recommends majoring in physics or electrical engineering.

3. Jobs won't disappear

• The idea that AI will erase 20% of jobs is "ridiculously stupid," LeCun says, adding some roles will disappear but new ones will emerge, as in past tech shifts.
• One reason LeCun is convinced that AI won't lead to mass unemployment is that historically it takes new technologies 15 years to achieve their promise in productivity gains.
• Meanwhile, AI agents will be able to help everyone with their job.

Zoom in: "Everyone is going to be a boss," he said, but a new kind of boss.

• You'll manage agents instead of people, meaning it will be more important to have a sense of strategy and direction.
• Skills in managing humans won't be as necessary, "if your staff is a bunch of AI systems."

Yes, but: AI may compress skill gaps: Entry-level workers improve more, while top performers see smaller gains.

Zoom out: LeCun argues this AI wave is not fundamentally different from past tech revolutions.

• "There is nothing qualitatively different between the previous technological revolutions and this one," he said. "It's just another set of tools that makes us more efficient."

LeCun will be honored as one of AI's "godfathers" at Liberty Science Center's annual Genius Gala.

• "In our society, we don't make heroes of scientists as much as we make heroes out of pop culture figures and musicians and athletes," Paul Hoffman, the center's CEO, tells Axios.
• "So part of our event is to really recognize scientists that are doing brilliant things that could help improve our world," he said.

The bottom line: Maybe the truth in the AI doom and hype cycle lies somewhere in the middle.

While AI was trained to write like humans, these large language models are eroding the unpredictability of our writing and shifting the way we talk.

The big picture: Researchers found that AI pushes users toward a more standardized speaking and writing style, reducing variations in sentence structure and vocabulary.

• The University of Southern California study analyzed scientific journals, local news articles and social media and found writing style diversity dropped sharply after ChatGPT's release.
• ChatGPT's favored words — such as "delve," "meticulous," "boast" and "comprehend" — are showing up more in everyday conversation, researchers at the Max Planck Institute for Human Development found after reviewing 740,249 hours of content.

The intrigue: "People get used to this idealized, very predictable form of language, and even people who are not using it, in order to have that sense of powerful, influential writing, they start writing more like LLMs," Morteza Dehghani, a USC professor who oversaw the study, tells Axios.

• Alex Mahadevan, chief AI instructor at the Poynter Institute for Media Studies, tells Axios that AI writing is noticeably "soulless" and "mediocre," even though it is grammatically correct. "There's no art in it."

Between the lines: Emily Bender, a linguist at the University of Washington, said she does her "very best not to read any synthetic text." But "oftentimes people will send me something and I won't know."

Keep reading.

• Apple raised the entry-level price of its Mac mini desktop to $799 from $599 amid a surge in demand for users wanting to run AI tools on it. (Bloomberg)
• AI is starting to replace the early-career grind of junior and summer associates, Big Law's most important classroom. (Axios)
• Bloomberg talked to a bunch of real-life Claudes and, yes, it's a weird time to be a human with that name.

You may be too old to go to summer camp, but now your mouse pointer can, thanks to "Cursor Camp."

• Yes, but: It's easy to lose track of your actual pointer, but pressing "escape" lets you escape.