In that spirit, here's me doing this about AI at Dropbox:
- Get consent from users and train a personalized AI on user files, to answer questions like, "Where is my meeting with Thiago?"
- Provide built-in support for number crunching on unstructured meeting notes: "How much revenue did we make, from last quarter's notes?"
- Automatically fine-tune a model using your own writing so that ChatGPT talks more like you
- Apply an AI to recognize security threats in your files("Hey you might have stored a secret in plaintext in file Y. Do you want to encrypt it?")
* find someone that will actually want to pay extra for their subscription on that.
What if it’s more about using customer data to Dropbox’s advantage? Combined with the telemetry you can’t turn off, they can mine the shit out of users.
I can see DropBox making some interesting features for their products around AI. That’s going to require manpower to do it. Manpower that’s probably a little nervous about staying at a company that just did layoffs.
This is genuinely blind to anything that's happened in AI over the last 6-12 months. The data that is the most valuable to AI is exactly all of the random, unstructured data (product requirements, decisioning documents, engineering designs, process documentation, etc etc) that has been shared across teams.
> I just can’t even come up with plausible uses for AI that meaningfully change the company enough to require a layoff.
requiring a layoff no, certainly me neither. There is definitely a bit of "compliment sandwich" happening in this letter. The layoffs were likely just a long time coming for a company that has been rather stagnant for quit some time.