Sam Altman says that bots are making social media feel ‘fake’ 

X -enthusiast and Reddit shareholder Sam Altman had a Epiphany on Monday: Bots made posts impossible to determine if posts were truly written on social media, he made it impossible PostThe

R/Claudcode came (and shared) some posts from Subradit, which appreciated the Open A. Codex. OpenAI Software launches programming services It takes on anthropic clad code in May.

Recently, the sub -dite self-proclaimed code has been so full of users’ posts that they have gone to Codex that a Reddit user has even joined: “Is it possible to switch to Codex without posting anything on RedD?”

This left Ultman was thinking that how many of these posts came from the true people. “I got the most weird experience of reading it: I assume that it is all fake/bot, though in this case I know that the codes growth is really strong and the trend here is real,” he acknowledged in x.

He then lived his argument live-analyzed. “I think a lot of things are going on here: The real people have raised the LLM-Speak joke, the highly online crowd has gone together in a very connected way, it has been too much of the hype cycle/We are very back ‘extremism, social platforms and other companies, which are related to the Creator.

To decide this to somewhat, he is accused of starting people like LLMS, though LLMs – opening -led – literally discovered to duplicate human communication, just to the bottom of the EM dash. And the Openai models Certainly trained in RedDWhere Ultman was a member of the board within 2022 and was Published as a large shareholder During the company’s IPO last year.

He has created a valid point that the final, always on social media users led the fondom in a odd way. Many groups can become disgust if they are printed by those who have spread frustrations to their brothers.

Sam also throws an excavation when social media sites and creators rely on the busyness for money. Enough enough.

However, Ultman then acknowledged that the posts of the Subredit Opener could be botted. It usually involves posts of persons or bots provided by the contestant or gives commendable denial to some third-degree contractor.

We have no evidence of astrotorfing (though it is possible). But we have seen how Openai Subradits Agency This is after the release of GPT 5.0. Many angry posts were voted instead of a wave of praise than the new model. People went to Reddit and X to complain about how it was burned through GPT personality Does not finish the jobThe

A day later Bump ReleaseUltman did a Reddit Ask me R/GPT is the session where he acknowledges the rollout issues and promised to change. GPT Subreddit has never fully recover the love of its previous level Users are still posting regularly They dislike the changes with GPT 5.0. What people are they? Or do they mean to Ultman, somehow fake?

Ultman currently flows, “The net effect is somehow AI Twitter/AI Reddit feels very fake that it didn’t really happen a year or two years ago.”

If this is true then whose fault? GPT Models have managed to become so good in writing that LLMs are not just on social media sites (which was always a bot problem) but became plague At school, JournalismAnd Court

Although we do not know how many Reddit posts are written by bot, or using LLMs, it is probably sufficient. Data Protection Agency Immpiva said More than half of all the Internet traffic in 2024 Was human, mainly because of the LLMs. X’s own Grock is your: “The correct numbers are not universal, but 2024 estimates offer a few million bots in X -in.”

Several camouflaces have suggested it Lamentation Was his first enthusiasm in marketing Openei rumorsThe In April, The Verge reported that the national project with X and Facebook was in the early stages. This product may or may not contain. Ultman nowadays social media may or may not have a spontaneous purpose to suggest that social media is too fake.

However, on the one hand, if the OpenAI plans a social network, what is the adversity that will be a bot -free region? And, enough fun, if it is reversed and forbidden people, the results will probably not be different. LLMs are not yet Hallucinet data, when researchers at the University of Amsterdam created a social network composed of bottles, they found out that Bots soon Consisting of cycle and echo chamber for themToo