Facebook's Offering OpenAI Employees 100 million dollars

According to Sam Altman Meta has been desperately trying to head hunt AI talent lately, to the point where some researchers and developers at OpenAI has been offered 100 million dollars in sign-on bonuses. This might sound absurd, until you realise how OpenAI is literally eating Facebook's lunch. To understand Meta's reasons here, let's imagine an average day in the life of Sara, a previously active user of Facebook.

Sara wakes up a Monday in the summer of 2025. She makes herself some coffee, and turns on her iPad to check the news. After she's done reading the news, she's got two choices.

• Login to Facebook
• Login to ChatGPT

If she goes to Facebook, about 50% of everyone commenting on her posts will disagree with her, sometimes to the point of that they're insulting her. This is "by design" because Facebook realised a long time ago that the more controversy and hostility users are experiencing on the platform, the more time they will spend on the platform, and the more ads each individual will be exposed for - Increasing Meta's revenue. If she engages in a public debate that's polarised from before, she might even get death threats, and even risk others "swatting" here, depending upon how polarised the debate is. In Norway, which is supposed to be a "peaceful country", there are recorded examples of people having reported hundreds of other users anonymously to the Child Protection Services (CPS) because of an argument over which political party is the best.

If she goes to ChatGPT, she will have a pleasent experience, where the LLM is sympathetic, and even support her with positive words - Also when she's wrong. Even though it hallucinates every now and then, it's nothing compared to all the Annunaki junk and crazy conspiracy crap she finds on Facebook - Some of it literally planted as malicious false information by vicious governments or criminal organisations, having their own agenda, and benefitting from having others quarrel over petty arguments. ChatGPT on the other hand will do its best to answer her questions, and because of the cost of running the LLM, they will not try to hold her on the platform for as long as possible, but will be quite content if she only spends 5 minutes on the platform each day.

In addition, Sara is paying for ChatGPT, so it's at least in theory supposed to be on her side - Because OpenAI has got synchronised incentives with their users, contrary to Facebook where the user literally is the product. This implies it's in OpenAI's best interest to help Sara anyways they can, to make sure she's continuing her subscription. While it's in Meta's best interest to make her life as miserable as possible, such that she becomes angry and triggered, ensuring she stays on the platform for as long as possible.

In addition, hundreds of millions of users perceive Facebook as their "single source of truth", which gives Meta enormous political power - Something clearly illustrated by how Facebook arguably have overturned dozens of governments, world wide. To the point some would claim it's also over turned the US Presidential election a couple of times.

With ChatGPT, the information she finds does not depend upon Sara spending hours in braindead arguments with the LLM, so naturally the information will be presented in a less "hostile manner", and arguably "less biased". This implies that Sara no longer looks at Facebook as her "source of truth", diminishing Zucerkberg's powers, if not for any other reasons than the fact that at least now she's got two sources for finding truth, which helps her balance her perspectives on the world.

Wrapping Up

Meta is desperate about AI, not because they want to create "the best LLM in the world" - Something that can be seen through the fact that they're releasing all their LLMs as open source. This is of course not because Zuckerberg wants to be a "nice guy", but rather because he wants OpenAI and similar initiatives to simply disappear, and the best way to ensure that, is to completely cannibalise the eco system, making it impossible to make money on LLMs, by open sourcing their own LLMs.

This implies that OpenAI is literally the first existential threat Facebook has ever experienced, and it might literally make Facebook go bankrupt over time - Because it changes habits - And Zuckerberg needs habits to stay exactly as they were back in 2022. And the paradox is that technically these companies are not competitors, so OpenAI is basically playing on a field where Meta can't even compete - Since OpenAI is not playing the "social game", but in fact a completely different game, that has the potential to "change the games we're playing". And this is simply not something Facebook can compete with. If you rather want to hear me talking about these events, you can watch the following video where I analyse the relationship between Facebook and ChatGPT.