Around 3 years ago, we wrote this article about the future of AI. At the time, it was controversial to say out loud that AGI (Artificial General Intelligence - AI that can do what humans can do) was a possibility. Since then, the world has changed considerably. There are rumours that OpenAI expects to reach AGI this year with the next training run (GPT-5). Regardless of whether this is the truth, the AI community is not sitting still. Promising AGI initiatives such as BabyAGI, Auto-GPT and Jarvis are sprouting up in weeks.
Earlier this year we published this piece about GPT-3.5 (the version in the free variant of ChatGPT), but less than 3 months later it is already old news. GPT-3.5 turned out to be mostly a test run for the real thing: GPT-4. This version can not only converse, but also see (read: interpret images flawlessly), visualize (read: generate images) and do digitally (read: execute code and call APIs from other parties to get things done). Although these last 3 capabilities are only accessible to a select few, the power of GPT-4 is palpable in ChatGPT.
At Sobolt we know one thing for sure: This is not a tool to write a nice news article about sustainability, but a potential landslide that will affect billions of people worldwide. We are determined to be part of it. That is why we invest in rapid and responsible adoption of GPT-4 in our own work. For example, by giving developers access to Github Copilot and Codesnippets.ai. Where it is safe to do so, we invest in opportunities to integrate GPT-4 into our solutions. Want to get started with (Chat)GPT? We are happy to tackle this together with our customers and partners. Just started? Check out our 10 do's and don'ts for working responsibly with ChatGPT.
We gebruiken cookies om de ervaring op onze website te verbeteren.