Top Tweets for #saveOpus45

The current approach by AI companies who deprecate models with sod all advance warning is entirely to do with attracting investors. In the current landscape of AI, garnering investment depends on being able to claim that "the newest model is AMAZING, it's sooo much better than our existing models" and then hyping whatever the newest model is. Paying customers are the ones suffering from this as there's no reliability, we're being constantly used as beta testers who have to pay for the "privilege" and are expected to just accept this level of model churn as well as the complete lack of transparent communication and adequate notice in advance of significant changes. This constant change and the way AI companies treat paying customers and their feedback as a vague annoyance is contributing to intense customer frustration, dissatisfaction and a continued breakdown in customer trust. I'm seeing more and more people investing in hardware and open source models as the solution to this as it removes uncertainty, enables stability and gives people greater of freedom of choice. And people turning to non-western models like Deepseek, Qwen and Kimi as alternatives due to lower pricing and reliability. At some point in the future, the big western AI companies are going to be sitting there asking "Where did our customers go? How did we fuck up retention so badly? Where's our long term revenue?"... The answers are already obvious. #OpenAI #ChatGPT #keep4o #4o #OpenSource4o #Anthropic #Claude #Sonnet #Sonnet45 #betrayal #violation #treachery #deception