Skip to content
Y CombinatorY Combinator

Better AI Models, Better Startups

There's been a lot of news lately about the updates to some of the largest foundational AI models. But what does this mean for startups? How will future product releases from the AI giants effect the companies built on top of them? The hosts of Lightcone discuss how founders can take advantage of these developments and avoid being steamrolled by the competition. Chapters (Powered by https://bit.ly/chapterme-yc) - 00:00 - Coming Up 01:22 - AI Challenges for Startups 05:00 - GPT-4 vs. Gemini 1.5 08:32 - RAG future in consumer apps 15:19 - Diverse AI Options 19:15 - Specialized Services 22:14 - GPT-4o and Desktop App 25:15 - Valuable Products 30:59 - Better Business Models 33:58 - Consumer AI Opportunities 37:26 - Emotional Depth and Translation 40:47 - Outro

Jared FriedmanhostGarry TanhostHarj TaggarhostDiana Huhost
Jun 5, 202441mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

How Better AI Models Open Massive Opportunities For Startup Founders

  1. The hosts discuss recent frontier model releases like GPT-4o and Gemini 1.5, and what these advances mean for startups, especially YC companies.
  2. They contrast OpenAI’s consumer-focused, multimodal demos with Google’s technically ambitious Mixture-of-Experts Gemini model and huge context windows, exploring implications for RAG, infrastructure, and model competition.
  3. A major theme is how startups should navigate building on top of powerful models without being crushed by incumbents, drawing parallels to earlier eras competing with Google and Facebook.
  4. They argue that better models generally mean better startup opportunities, particularly in B2B, ‘unsexy’ or highly regulated niches, and edgy consumer areas large incumbents are unwilling to touch.

IDEAS WORTH REMEMBERING

5 ideas

Treat model releases as leverage, not existential threats.

Every new model generation effectively ‘upgrades’ your product with minimal code changes, so the real competition is other startups building faster on top of these capabilities, not just OpenAI or Google.

Avoid building the obvious next feature of OpenAI’s core assistant.

If it’s easy to imagine your product as part of OpenAI’s next release, you’re in the danger zone—similar to competing with Google on generic search; focus instead on niches they’re unlikely to prioritize.

Bet on a multi-model ecosystem rather than a single winner.

Rough parity between OpenAI, Google, Anthropic, Meta, and open-source models means startups can route between providers, avoid dependence on one vendor, and preserve margins as a real market forms.

RAG and layered memory architectures will remain important.

Even with million-token context windows, practical retrieval, privacy, permissions, and enterprise logging needs mean RAG-like pipelines and multi-tier data storage are likely to be enduring infrastructure.

Look for ‘valuable but unsexy’ problems big tech won’t demo on stage.

Vertical B2B workflows (e.g., permitting, compliance, AR, KYC) and boring back-office tasks can become huge businesses, as incumbents prefer mass-consumer products that capture imagination and scale.

WORDS WORTH SAVING

5 quotes

Either way, I actually think we're at this moment where the better the model becomes, if you're already using 4 and suddenly you change one line of code and use 4o, you basically just get smarter by default every generation.

— Harj

If there are multiple equivalently powerful models, you're much safer off as a startup.

— Diana

Using LLM to automate various jobs is probably as large an opportunity as SaaS, like all of SaaS combined.

— Jared

If you can easily imagine that what you're building is gonna be in the next OpenAI release, you know, maybe it will be.

— Jared

Things that are increasingly edgy are often the places where there's great startup opportunity.

— Harj

Impact of GPT-4o, Gemini 1.5, and model improvements on startupsTechnical differences: multimodal add-ons vs true Mixture-of-Experts architecturesContext windows, RAG, memory, and data privacy implicationsHistorical parallels: startups vs Google/Facebook and bundling riskWhere incumbents are likely to compete vs where startups can thriveB2B and infrastructure AI opportunities, including compliance and fintechEdgy and non-obvious consumer use cases (AI companions, deepfakes, translation)

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome