There’s never been a better time to start an AI company. Not just because there are new ideas, but because the tech finally makes old ones actually work.
On the Lightcone, Garry, Harj, Diana and Jared talk through the kinds of startups that are suddenly viable thanks to LLMs—from full-stack law firms to personalized tutors to recruiting platforms that can finally scale. They share the patterns they’re seeing, the ideas they’re excited about, and what it means to live at the edge of the future, where breakthroughs often look like second chances.
If you’ve been waiting for the right moment to build, this is it.
Apply to Y Combinator: https://ycombinator.com/apply
Work at a startup: https://workatastartup.com
Chapters (Powered by https://chapterme.co/) –
0:00 Intro
00:41 What startup ideas could not work before AI?
06:06 Technical screening products
07:35 Truly personalized education tools
09:48 Do better products automatically get better distribution?
14:41 Moats
16:08 The need for platform neutrality
17:40 Big Tech and AI
23:24 AI horseless carriages
25:14 Gross margins
30:03 Full stack companies
32:30 ML ops
37:14 Updated startup advice for the AI age
40:19 Outro
Audio trasltate spanish please
Startup ideas that you can go do to keep our savings working for us! Weeeee.
If they're grading assignments, we'll that kind of takes the teacher OUT OF IT.
Oops, another industry slaughtered…
Always ask why? Create from source 👌👌👌
I am living in Perth. Last week I found out bolt and have generated two tools that I think they were needed for my own needs. I have literally no knowledge what so ever related to tech industry. I am a tiler. It is crazy that stream of ideas keep flowing into my brain at the moment like they said at the end of the video. Thinking to take an online course to learn Javascript and how to build AI agent. I feel like I am so far behind and at the same time no one around me has the same interest. Im glad that I found your channel. 😆
Hit home on this one. My first legitimate business plan was in 2003 as a Sr in HS. It was called "Store To Your Door", a mutually beneficial efficiency/convenience solution. In those days majority of my time was spent skateboarding relentlessly all throughout the Twin Cities. One particular day this thought hit me like a ton of bricks, with urgency too. I couldn't make sense of how taxi/cab drivers were all over the city 24/7 with such a massive amount of downtime. Literally reading novels in shopping centers and restaurant parking lots for sometimes hours. So I started deploying forward and interviewing any driver I would run into and pitching the concept of delivering goods In their downtime. Long story short, the drivers loved it, everyone else did their best to make sure I knew it was a ridiculous concept. Turns out I was just way too early. Cell phones were mostly to play snake at that time, Internet was still a possible fad to many and payments were rough, cash was still king. I still feel like i could have been on to something, lol. Timing is everything. Still building, still looking for collaborators, still deploying forward.
AI meeting summaries include people that never were in the meeting apart from it not even understanding a ton of what has been said.
But sure, please tell me about how amazing AI is.
As an EdTech researcher and TA, I'm honestly perplexed that it seems broadly acceptable for faculty to openly dislike grading. Grading isn't just paperwork, it’s one of the best ways for teachers to see if their teaching actually work.
I think a lot of professors dislike grading because either they haven't created assignments they find meaningful, or they see their job as just delivering content, leaving learning entirely up to students. This means grading often becomes an afterthought or something delegated fully to TAs or AI.
Don’t get me wrong, I appreciate being a TA and the financial support it brings as a grad student. But teaching works best when professors and TAs collaborate, not when grading is completely handed off. To me, this issue reflects a deeper question of work ethics and commitment to teaching, rather than merely a productivity concern.
Note: I checked out the Edexia demo video, and it seems they're developing with a teacher-AI collaboration mindset instead of just focusing on productivity increases, which is very encouraging.
Remember that team that built a product for middle managers to micromanage then crack down on wage slaves that weren't working hard enough
What? Gemini in emails is not good??
Intelligence, context window, TPUs, GPUs. Is there a nice book for beginners to understand some of these items fundamentally?