When Open AI released Chat GPT in 2022, the whole world stood up shocked at what AI could do. The market poured in and so did the competitors.
Since then Antropic and Google trailed close behind. Grok came and … existed. While companies like Meta, Mistral and many more tried to be relevant in the race.
Later on the Chinese models started catching up, where DeepSeek shocked everyone in January 2025 with its open-weight reasoning model. This has since led to providers like DeepShot AI (Kimi), Alibaba and many more fighting for the lead in the various AI benchmarks.
We’re now at a point where the AI analysis site lists 57 different model providers. Thats at least 57 companies generating one or more models, most of which have the same goals and purpose. On top of this none of these AI divisions have worked out a healthy profit model yet, therefore they need to project massive growth, but how?
Well they’ve got to keep proving they’re the next hot thing.
In 2022 every investor would give OpenAI an arm and a leg to get a piece of the pie. Then for years a massive amount of competitors have been funded by this AI hype. All kept alive through announcements of (dreams of) vanishing jobs and revolutionising industries.
Now, the space is so competitive, every company is in a fight to stay relevant. Each week its like children fighting for their parents attention. All leading to the past two weeks, which has been everyone’s favourite dick measuring contest yet.
Now this is not to downplay the impressive (unintentional) hack by Open AI on Hugging Face. This is another moment where we can say, “You know what, maybe Open AI isn’t falling behind Anthropic for software”. But the other AI labs couldn’t let Open AI get the last laugh.
Anthropic basically did an @OpenAI Blog Post to say, look were super smart and scary and can accidentally hack too. Come on, that just sounds a little desperate. But luckily for Dario Zuck said hold my beer, swooping in with the classic “Yeah I also did a hack by accident to another company, they’re super smart and pretty but go to a different school”. Has anyone even heard of a model Meta has created in the last 2 to 3 years? It’s so embarrassing feeling the need to jump on the band wagon here when, no one believes them. It makes them look weaker.
Now accidentally performing very serious, illegal actions is a boast to prove the AI’s capability. Are we eventually going to be at the point of having all these AI providers creating models which are going around hacking companies? That becomes the expected capability now? But that kind of power in the hands of many companies?
And from that point, what will be the next sign that a provider is truly the market leader? They give you the nuclear codes? They make planes fall out of the sky? They will convince you ex to get back with you?
The AI mania is entering its 5th year soon. The competition is very high, the financial investments are very high, the providers have to start pulling off greater and greater stunts in order to impress the market. And either some of these providers will fail due to too much competition, or we end up with many companies creating dangerous AI models.