It: I am AGI. Ask me anything.
Me: How many Tesla’s were registered in July of last year?
It: thinking… searching Bing… searching Google…
It: While the data isn’t clear, here are a bunch of things humans wrote about Tesla in July 2023 for other humans.
This seems to be the mode we are operating in currently. If the LLM and SLM we are employing needs more information to formulate a response, it now has access to the vast oceans of human created content on the internet.
This seems extremely limiting as most human created content was written with human audiences in mind. Barring a small percentage of words penned, most writings were created with human audiences in mind and this fails to address the needs of a new audience of content consumers.
At Machine Generated, we believe that machines are going to become the most voracious consumers of content going forward. They can and will want to consume multiple orders of magnitude more content than was created since the first caveman put his handprint on a rock.
The limits on how much content humans can produce are a function of the number of humans creating content and the time they dedicated to creating it. What if we brought the marginal cost of creating an additional piece of content to near-zero? What does this do to the amount of content we create and the audiences we can now cater to?
Having machines create the content that is consumed by machines is the answer but not without its challenges. From data source quality to inference to responses, all aspects of machine generated content production needs to be orchestrated with precision to ensure high-fidelity outputs.
At Machine Generated we are tackling these challenges and are excited to find ways to create content for these new consumers of content.