AI explainer videos from a former Google Research engineer, now startup founder. Expect American optimism with a solid dose of Eastern European cynicism.
Find (free) supplemental materials (slide decks, paper reading lists, Colab notebooks) on Patreon π www.patreon.com/c/JuliaTurc
Julia Turc
I'm curious: How are we feeling about sponsored content?
When I started the channel, I never imagined I'd do sponsorships (if I were optimizing for money, there would be a thousand better ways to get it than YouTube). What's enticing about sponsors is access to their talent; they often share things that are not public and give me very useful guidance around what matters and what doesn't for a topic. Granted, I could still work with these companies and not charge them anything (therefore keep 100% independence and credibility), but logistically that just makes people flakey.
While I'm truly enjoying building relationships with the best people in AI, I also don't want to alienate the audience. So I want to hear from you! π
1 month ago (edited) | [YT] | 91
View 43 replies
Julia Turc
Working on a video about specialized LLM chips (TPUs, Groq, Cerebras, Taalas). Let me know your questions! π
2 months ago | [YT] | 291
View 62 replies
Julia Turc
I'm thinking of trying out some video formats in addition to my standard explainers. All of these are interesting to me for different reasons. Which ones would you enjoy most?
2 months ago | [YT] | 39
View 22 replies
Julia Turc
Unusual question here. Why do people still watch my channel (and, more broadly, technical explainers) given that AI is such a good tutor? I have my own answer for watching 3B1B / Welch Labs / etc, but I'm just n=1. Curious what your answer is! P.S. No need for flattery, I'm just trying to figure out the larger trend / sentiment.
3 months ago (edited) | [YT] | 125
View 121 replies
Julia Turc
For my next video on full duplex speech models, I might be about to interview NVIDIA about PersonaPlex. research.nvidia.com/labs/adlr/personaplex/ LMK your questions! π
3 months ago (edited) | [YT] | 47
View 4 replies
Julia Turc
I've made quite a few videos about language, image & video, so... audio shall be next! I'm not sure where to take it yet (ASR, TTS, realtime, etc). Which aspects of audio models are you most curious about?
4 months ago | [YT] | 154
View 49 replies
Julia Turc
Currently working on a video about world models. Let me know your questions! π
5 months ago | [YT] | 119
View 24 replies
Julia Turc
Iβm seeing some constructive comments on my videos suggesting that the pace is too rapid. Iβm trying to gauge whether that is a universal feeling. I usually go as fast as necessary to squeeze the content in 20-25 minutes, but perhaps thatβs not the best strategy. What would you prefer?
5 months ago | [YT] | 43
View 24 replies
Julia Turc
The Flow Matching video is finally up! What would you prefer to see next?
5 months ago | [YT] | 42
View 12 replies
Julia Turc
I had a lot of fun nerding out with Subham Sahoo, the researcher who pioneered the sub-field of *discrete* diffusion models.
This interview format might confuse the algorithm for my YouTube channel, so I published it on Patreon instead (free, no sign-in required):
www.patreon.com/posts/interview-with-155096409
5 months ago | [YT] | 52
View 4 replies
Load more