Interviewing Bilawal Sidhu: AI, VR, AR, The TED AI Show, Google, A16Z Scout, 1.4m on YouTube + TikTok

​Bilawal Sidhu is the host of the TED AI Show, a company scout for venture capital firm A16Z, a creator on YouTube TikTok with 1.4M subs, used to work at Google, and has a focus on AI, AR & VR:

​

🎧 Prefer listening to this...here's a podcast version! 🎧

​

​

(0:00) Bilawal is host of the TED AI Show, a company scout for A16Z, creator on YouTube and TikTok 1.4M+ subs, worked at Google, and has a focus on AI, AR & VR.

⬇

​

(0:30) Who is making money with AI? Who is not?

⬇

​

(0:50) The people making the real money in AI is the hyper-scalers like NVIDIA. Selling the AI services so you can offset the cost of training.

⬇

​

(1:20) MidJourney was doing $200m/yr in revenue, but most of the money is plowed back into training gigantic AI models.

⬇

​

(2:25) It used to be trendy to train your own AI model, but it’s getting so competitive and expensive that Microsoft and Google and others are literally buying into nuclear reactors in order to just get enough ENERGY to train these giant models.

⬇

​

(2:35) In the “ChatGPT Wrapper Era” all the AI companies were making money immediately, but ChatGPT dropped a huge bomb on that business model.

⬇

​

(3:30) This mirrors the trend in mobile apps, like the Flashlight App which just got absorbed into a feature of all phones.

⬇

​

(4:25) Where is the hype in AI and where is it real? I personally thought writing blog posts with AI would be huge, but in reality it’s not what people need anymore. LLM’s are huge right now for productivity needs like summarizing and automatic transcript

⬇

​

(6:00) On the visual creation side it’s getting amazing: Ideogram.ai is phenomenal and even gets writing correct (versus other AI's that still can’t write text in images). Flux as well through BlackForrest Labs. Even layering has been solved.

⬇

​

(7:45) The hardest thing for most people making content is making images. I used Photoshop and was 10x ahead of people, now the AI stuff is levelling the playing field.

⬇

​

(9:13) Video is still not completely there yet. We see cool demos but it’s still a little weird looking.

⬇

​

(10:05) Rowan Cheung (biggest AI newsletter) has automated his short videos with AI so he can pump them out insanely fast. Everything even video is automated. People don’t really care if it’s AI if it’s valuable.

⬇

​

(11:03) People who are 30+ might care if it’s written by AI for a while, but eventually it seems no one cares so long as it’s helpful and accurate.

⬇

​

(12:58) VR keeps popping up then going away (seemingly), but it’s starting to all come together as the form factor approaches lightweight AR glasses. When it can replicate having a full desktop setup it’ll be very helpful.

⬇

​

(15:46) Sometimes I daydream about technology, and if you could have ChatGPT-linked glasses that just display text, you could rule the world as the fastest/smartest person.

⬇

​

(16:45) Bilawal started off working on Google Glass, and it was primitive by today’s standard, and before LLM’s were around. He was surprised Apple went full immersion vs starting off with some lightweight glasses that did basic text.

⬇

​

(19:28) I’ve been into VR for a while, but stopped using it because it was antisocial in a way, you can’t really connect with others…until Apple Vision Pro. They need a few more years to go, but they did things to where it truly “mixed” reality and virtual

⬇

​

(23:09) The next version of Tinder will be pop-ups over reality showing who is single in a bar.

⬇

​

(23:30) When I get older and get left behind in technology, it’ll probably be because kids live natively in a mixed virtual and real world and have more experience inside of it.

⬇

​

(24:54) Location services for all your friends are pretty cool and not creepy as once thought. It’s actually really helpful and fun. For some kids I’ve seen their parents give them an Apple Watch but not a phone so they can have all the benefits without all the distraction.

⬇

​

(26:22) “It sucks we use a slab of glass to mediate reality. It’s a distracting form factor.”

⬇

​

(27:00) Mapping the world. The goal is to create a 1-to-1 digital twin of the world.

⬇

​

(29:22) I’ve used Waymo (driverless cars) 40 times in a month and it maps out it’s whole world digitally and you can see it while driving.

⬇

​

(31:12) “What are our top social platforms right now?” X and LinkedIn. Biggest account is TikTok but it’s lower engagement. Uses X and LinkedIn like a journal, it’s the least friction way to document insights, and also the people I care about are here. X: All the startup founders, researchers,VC’s are here. LinkedIn: Managerial class and older people not on other socials.YouTube / TikTok: The sandbox and playground to try out new tools

⬇

​

(33:15) What platforms “impresses” people the most? People are more impressed with YouTube than TikTok. TikTok and YouTube are mostly entertainment based or “fast food education.” On X the people are finding the latest and greatest and are more curious. Established old money people are more impressed with X, YouTube, the TED affiliation. Often times the “impressing” is from a combination of all previous credentials.

⬇

​

(36:35) What is the social media “content strategy?” He was avoided just finding & reporting the latest news, it’s difficult and a treadmill, and others do it better. Try to cover the stuff no one else is because they don’t have the domain knowledge.

⬇

​

(40:09) Bilawal will be the technology curator at the main TED conference. The theme will be “Humanity Re-imagined” in April. TED was the original “cool science videos” on the internet. They are going back to those roots.

⬇

​

(42:28) The flow of podcasting….what comes from it? Used as a top of funnel in a sense, and so far the TED AI Podcast has been all audio.

⬇

​

(45:01) Creator Economy stuff. There’s now a lot of attention, and a lot of good people competing…so where is this all going? You still have to have “other income” to pay for podcasting. YouTube is still the best passive monetization, X is starting now, but you still have to make full time income from elsewhere. The generative AI thing will be like a nuclear bomb on this industry and we don’t know how it will play out. It might just be whoever puts out a TON of quick consumption content.

⬇

​

(50:07) Many people put out content but never get paid for it directly.

⬇

​

(50:10) The Vsauce story how he made amazing videos, but despite millions of views per video, still not enough to sustain it. If a better pay structure was around I think it would’ve been better for the world for him to keep making amazing videos.

⬇

​

(52:09) Hopefully X can “tip” all your creators. Streamers have found this out a while ago making money from their streams.

⬇

​

(52:59) The way you monetize now as a creator is build that island of influence then offer some service or product or course or startup. This is also happening in the startup world as people are looking not as much for money but for distribution.

⬇

​

(54:14) A lot of creators in the AI space are making their money through courses on Kajabi or Maven, not directly from the content. Bilawal tried cohorts, but was burned out by 2 cohorts.

⬇

​

(56:56) Google saved the web with AdSense. I had “Neville’s Cool Car Archive.” Just putting up ads on your website made it profitable to get traffic. Hopefully there’s a new thing that does this with social content.

⬇

​

(58:21) “With every video I upload, I feel like I’m training my AI replacement.”

⬇

​

(59:54) What is your plan for 2025? I am 1 year and 8 months into doing this full time. Definitely a digital products approach and a book. Maybe something else not attached to the name as well. Advising + investing in startup.

⬇

​

(1:02:26) I have seen some of the best podcast monetization is not taking sponsorships, but rather something like MyFirstMillion which got bought by HubSpot, and HubSpot writes a big check every month to keep doing, and HubSpot “buys” all the ad space for their own product.

⬇

​

(1:04:28) Starting a fund is a great way just to even meet and talk to amazing founders. It’s the closest way to scratching the “build a product” itch.

⬇

​

(1:05:49) Does AI kill jobs? There’s a constant cycle of job destruction and job creation. This is just speculation over a decade (which is hard): Overall there’s a downward pressure on jobs, or a contraction of them. The “Generalist Specialist” - If AI is good as persona specific tasks, you still need the conductor of the symphony or a human in the loop. So the people good at a bunch of different things and linking them together will be powerful. A conductor knows HOW to play instruments, but is more effective by having others play them together. Data centers might turn into boring cubes, consuming large percentages of energy GDP, people will need to keep these going.

⬇

​

(1:08:26) As robotics lags behind the world of software, a lot of jobs will move to the physical world. We’re already seeing a trend where students are doing firemen or plumbers rather than going to a Big 4 firm, and that might be good for humans going back to working with hands.

⬇

​

(1:09:14) A lot of companies are going to be more effective with smaller teams. Competition on YouTube and other digital things will go up.

⬇

​

(1:10:59) When I was in college my parents preferred if I had a job at a big company because that was the way to succeed back then…but the internet happened and now a new way was available. But the next generation might think a makeup artist or barista is the bigger opportunity than doctor.

⬇

​

(1:12:54) There could potentially be a world where AI gets so good that we could merge with machines and live forever, and “making money” might not be that important.

⬇

​

(1:13:29) Could also be another portion of society that moves backwards like Amish to go back to all-human without computers.

⬇

​

(1:14:59) In my lifetime, what used to take 100 people takes 1. For about $3,000/mo I can personally put out more content and distribute it further than a local TV station, radio station, and newspaper combined. That was 100’s of people working…now it’s 1. So what will the next thing be that takes 100 people, that 1 person can do? Make a full-length movie? Make a giant company? Mainstream news?

⬇

​

(1:15:36) Jason Calacanis was talking about “Generation Toolbelt.” At the moment it’s better to be white collar than blue collar. The difference with this generation is they might be able to be an electrician for 6 months then go travel for 6 months. This might actually be good because we don’t have enough people for these jobs.

​

​

​You can see the entire interview, pod, and extended notes here →​
Hope you enjoyed this interview,
Neville Medhora

​

​