7 Comments

Great article! I found Invideo AI to be quite a gamechanger. I think content like that will absolutely FLOOD Youtube. The voiceover is just believable enough and the stock footage is just relevant enough that it will be possible to generate seemingly authoritative videos on any subject.

Also, I'd love to see you post a deep dive on Open Interpreter. Powerful stuff there, but the limitless system access seems like a bit of a privacy concern.

Expand full comment

Agreed! I had the exact same thought as I was testing it.

There are already tons of YouTube channels that release nothing but hastily assembled news items or trivia facts voiced by uninspired monotone robots. Now producing something like that could literally take a few clicks and a single sentence.

But I kind of hope that most people will quickly learn to tune these out as noise, so the algorithm continues to reward high-quality, well-researched videos. Time will tell, I guess.

Yeah, the blanket access is what makes me hesitant to install the Open Interpreter. Not as much the privacy aspect, but its ability to manipulate most of the settings, etc. Having said that, if there's a polished and tested Desktop version (I'm on the waitlist), I might just take it for a proper spin.

Good to see you here in the comments again, it's been a while!

Expand full comment

I don't know about you, but I do a fair amount of Zoom meetings. An accurate transcript would be incredibly helpful.

Expand full comment

Yeah having someone transcribe a meeting, assign speakers to each statement, summarize the main action points and generate a complete minutes of meeting is a huge value add for sure!

Expand full comment

Have you experimented with this yet? I'd like to understand the limits if possible.

Expand full comment

I have not tried any of the collaborative meeting tools as they require inviting a bot that listens in to your meeting. I'm sure companies need to consider data privacy issues with that, etc.

But I did try Yoodli (I wrote about that here: https://www.whytryai.com/p/10x-ai-5-draggan-microsoft-build-generative-fill), since it only listens to your side of the call. I found its transcription and analysis capabilities to be quite great.

So I think the hurdle here is mainly legal and security-related, rather than technological.

And you have so many companies in this space:

https://supernormal.com/

https://meeting.ai/

...and probably dozens of others.

Expand full comment

Good points on the legal hurdles being the main obstacle. I think that's going to be the case for more and more things - we're seeing liftoff in a lot of disruptive areas (like this one).

Expand full comment