pz

This approach makes a lot of sense. Advertising is a marketplace and this is a great way to bootstrap advertising inventory. Its inevitable they will allow advertisers to manage ad spend directly through OpenAI but right now the product is too new to capture meaningful ad budget. This way they can begin testing delivery and develop proof points around ROI and build towards larger ad spend directly.

Yes, we've discussed that, it makes sense IMO

> Being able to tag chats with keywords would be nice. Being able to pull chats into docs easily would be nice. Being able to pull chats into more than one doc -- not just move, but tag/reference/copy. Global tags (with ACLs) as well as team and personal tags (also with ACLs) would be fantastic. Don't forget read and access ACLs, not just write ACLs.

Tagging is another organizational feature that is on our roadmap and has always made sense to me. Right now our organizational model is primarily hierarchical which has obvious limitations.

RE: email integration - We have a pretty robust email integration right now. Neville was always insistent that we shouldn't force people into the app to have a conversation, especially for one-off collaborators who get looped into a chat. They can stay blissfully ignorant of the fact that the conversation is actually happening on emdash if it suits them.

This is in the works. We have a prototype integration with Github that supports the deep search use case you mentioned. My favorite anecdote from this was a few weeks ago I was revisiting some work on transactional support in our controller framework and asked assistant "A few months ago I was working on transaction support and had to revert my change. Can you remind me what happened?" and it spit back my original PR, the reversion, and a deep link into the standup video meeting where we discussed the issue. For me, that was the magic moment where I knew this could be something much more than just a chat client.

This is good feedback, thanks. We packed a lot into the product so far and could have spent more time focusing the media assets to distill things better.

Early adopters range in size from 2 people to ~20. As you said, the Catch-22 for larger teams usually have established tool stacks so the (operational) switching cost is prohibitive.

FWIW we are a team of 5 and already find the feature set useful (we're biased, of course). I expect that ~5 is the threshold the organizational and search features become invaluable.

Amen. We've migrated teams from Workplace who really appreciated our hierarchical discussion model

Thanks! One of the benefits of dogfooding our product everyday is that we invest a lot in working out the everyday kinks in addition to the marquee features.

I should introduce the rest of the crew. We're a team of builders who have worked together for the better part of the last decade. Neville and I were early engineers on Facebook Ads. Neville later founded Rimeto, which was acquired by Slack. Fred has held various roles driving growth and was an early employee at Facebook, Doordash, and Rimeto. Nick jumped into startups right out of college, joining Rimeto and now diving back in with us at emdash.

In particular, I appreciate that you assign value to the consequences and not the decision itself. Anytime junior engineers on my team would complain about "shitty code" I'd assure them that someone would be complaining about their code in a few years.

Having the context or, better yet, responsibility for the past decisions is great for developing a pragmatic approach to software design AND empathy for other software engineers.

"Caring about the economy" and "analyzing returns" are not mutually exclusive.

its more likely he was mocking doomsayer articles like this one