Function calls are definitely one part of adding context, but lots of others that codegen tools are probably missing. Have you guys considered post-mortems, pagerduty outputs, slack threads focused on specific issues, etc.?
A private equity firm once asked the consulting firm I worked at to do a due diligence of a cold chain logistics company. It was ~5 years ago so perhaps things have changed, but some interesting tidbits:
+ Makes sense that this startup is starting with medical applications -- vaccines going bad because of poor refrigeration is a well-studied problem
+ I recall one cold-chain company being well known for its back-up battery because medical products are often stopped in customs, and the boxes cannot be plugged in. So you want boxes that have backup batteries that can remain unplugged for a few days while going through customs checks
+ High end seafood is another big application -- the company we were looking at started transporting lobster before moving upmarket
At the time there was a PE blitz to get into cold chain -- it had a lot of factors that they look for -- high margin, recession proof (at least the medical applications), etc.
Love it. One of the things we're focused on is getting better anatomical movement out of existing models. These models are great with animations like waves, cars, trains, flames, etc., but they struggle with people outside of basic movements. We're optimistic about future models, and we still think there are interesting, funny, exciting stories you can tell with what's available today!
What do you use it for? I haven't found a great use for it myself (outside of generating assets for landing pages / apps, where it's really really good). But I have seen endless subreddits / instagram pages dedicated to various forms of AI content, so it seems lots of people are using it for fun?
Have you thought about making more structured objectives? Shameless plug: I built a simple French/Spanish chat app and the #1 piece of feedback I got was that it's a bit aimless without any objectives.
I built out ~20 scenarios that would give users objectives like buying train tickets or ordering coffee. Now that's the most used part of the app (not the generic chat bot). As far as ads go, App Store ads are not bad and can't be ad-blocked, so that's an option.
Hey Peter, that's a great point! I've listed my thoughts below:
A couple things:
1. From a compliance / risk point of view, the customers we've spoke to treat AI-enabled tools differently from standard enterprise products. The main reason is because the standard terms for free-tier usage specifies that model developers might use your inputs to train future iterations unless a user opts out. For that reason, a single third party that doesn't do anything with your data is better than the various AI tools that employees might use (and if you don't have an enterprise license with one provider like ChatGPT, that won't stop your employees from using free alternatives).
2. When it comes to sensitive data, there are different tiers. Companies like to keep their customer lists private, for example, but many tools in the sales stack rely on reading that data out of Salesforce as long as those vendors pass a security/compliance audit. Private keys are another tier of sensitivity, and we absolutely don't expect our users to share those. They can, however flag other text strings like "key=" or "API_KEY" or "username"