When Everyone Can Build AI

Last week, I repeatedly found myself rewinding a podcast that fundamentally shifted how I think about AI's impact on business (especially the creative industries). The Dec 6 episode of podcast Invest Like The Best, featured venture capitalist Chetan Puttagunta and a renowned anonymous tech investor known as "Modest Proposal." What caught my attention wasn't just their technical insights, but the implications for creative businesses like ours (and no doubt many of yours). After several listens and discussions with our team, I realized why this shifted my point of view about what the impact of open source models like Llama – and how small teams may win the day.  

The Revolution of Small Teams 

The podcast that sparked this reflection revealed something that fundamentally changes the dynamics of how technology advances. For the first time, small teams of 2-5 people can compete at AI's frontier in ways previously reserved for tech giants and well-funded labs. This isn't just about democratizing AI development - it represents a potential restructuring of how technology innovation happens.

The story begins with a significant shift in AI development. All the major AI labs hit what Chetan Puttagunta called a "plateauing effect" in their traditional approach to improvement. The old playbook - throwing billions at bigger models - isn't yielding the same returns. As Modest Proposal explained, "All of the world's knowledge effectively had been tokenized and had been digested by these models."

This limitation has forced a fundamental shift from massive pre-training efforts to what's called "test-time compute" - making existing models smarter through better reasoning and problem-solving approaches. Combined with Meta's release of their open-source Llama model, this shift creates a new reality: innovation no longer requires massive infrastructure or investment.

The implications extend far beyond just making AI development cheaper. When small teams can compete at the frontier, it changes:

The Pace of Innovation: We're moving from a world of annual or semi-annual improvements from major labs to potentially weekly or daily advances from hundreds of small teams working on specific problems. As Puttagunta noted, "In the last six weeks of time, the number of teams we've met here at Benchmark that are two to five people...is extraordinary."

Market Dynamics: The traditional moats of massive compute resources and proprietary data become less relevant. Success will increasingly depend on creativity, speed, and precise problem-solving rather than raw capital. This fundamentally changes the competitive landscape across industries.

Business Models: When the cost to build sophisticated AI applications drops by orders of magnitude, entire business models that weren't previously viable suddenly become possible. Solutions that might have required $20 million in investment can now be attempted with $200,000.

Distribution of Power: This shift potentially redistributes power from large, well-funded organizations to smaller, more nimble teams. It's reminiscent of how cloud computing democratized startup formation, but potentially even more dramatic in its implications.

This dynamic creates both urgency and opportunity. While many businesses focus on immediate challenges, the underlying technology landscape is shifting dramatically. Companies can now build sophisticated AI capabilities with small teams and modest budgets. What previously required millions in investment can often be accomplished for thousands.

But there's a darker side to this accessibility. The same forces making it easier for us to innovate also lower the barriers for competition. As Modest Proposal noted in the podcast, "If you can make something for $5,000 that costs two million, you have a fiduciary duty to your shareholders to find the best product for the lowest price." This isn't just about cost savings - it's about survival.

Consider what this means for creative businesses. If you're a post-production house with specialized expertise in color correction, what happens when small teams can build AI tools that match your capabilities? If you're a talent agency with proprietary processes for contract negotiation, how do you respond when competitors can automate those processes at a fraction of the cost?

Reimagining Creative Industries

The true opportunity isn't just in automating existing processes - it's in reimagining how creative industries could work. We're entering an era where:

Production timelines that once took months might take days. Think of a small production team using AI to generate initial edits, automate color correction, and handle basic sound design, freeing up human experts for the truly creative decisions.

Deal velocity will increase exponentially. When contracts can be automatically translated, negotiations streamlined, and terms standardized through AI assistance, the pace of business naturally accelerates.

Creative exploration becomes more accessible. Small teams can experiment with ideas that would have been prohibitively expensive to test just months ago.

The infrastructure to handle this acceleration - the tools, platforms, and processes - largely doesn't exist yet. That's where we think the real opportunity lies and what underlies Basa.

Starting Smart, Scaling Fast

For creative businesses navigating this shift, the first step is building understanding. This doesn't mean becoming an AI expert, but it does mean learning enough about current capabilities to make informed decisions. Start by identifying specific problems in your workflow that AI might help solve.

In our case at Basa, we're beginning with targeted experiments in contract analysis. We're not trying to build a comprehensive AI solution overnight, but rather learning through focused projects that deliver immediate value while building our expertise.

The key is finding the right balance between immediate needs and future capabilities. Look for projects that:

• Solve current pain points while building toward larger opportunities
• Can be accomplished with small, focused teams
• Deliver clear value even if the technology isn't perfect
• Help build institutional knowledge about AI capabilities

Meeting Markets Where They Are

Something crucial I've learned from thousands of customer conversations in the creative industries: most businesses aren't ready to talk about AI. Our clients - from major media companies to indie production houses - are focused on solving immediate problems and meeting this quarter's goals. When I mention AI capabilities, I often get blank stares or polite nods. Not because they don't understand the potential, but because they're wrestling with urgent challenges that feel more pressing. 

A major talent agency executive recently told me, "We know AI is coming, but right now we just need to get these hundred deals closed before the end of the quarter." A production company head said, "AI sounds great, but we can barely keep up with our current workflow." These aren't people resistant to change - they're pragmatists dealing with immediate pressures.

This reality has shaped our own approach at Basa. While we're actively exploring how AI can enhance our internal capabilities, you won't see AI mentioned in our customer-facing materials or demos. We've learned that solving immediate pain points - streamlining deals, reducing administrative overhead, accelerating negotiations - resonates more than promising AI innovation. The market will catch up to AI's potential, but for now, we need to meet people where they are.

A New Kind of Balance

As we navigate this transition, companies face an interesting challenge: balancing internal AI innovation with external market readiness. At Basa, we're building AI capabilities to enhance our operations and prepare for the future, while focusing our customer conversations on immediate, practical value. This dual approach - innovating internally while maintaining market alignment externally - might become increasingly common across industries.

The key is understanding that adopting AI internally and selling AI-powered solutions externally are distinct challenges that may move at different speeds. Your organization might be ready to leverage AI for efficiency and innovation, but your market might need time to catch up. This isn't about hiding AI capabilities - it's about timing their introduction thoughtfully.

The shift we're witnessing isn't just about technology - it's about opportunity. The tools to build sophisticated AI applications are becoming more accessible every day. The question isn't whether to engage with AI, but how to do so in a way that creates lasting value for our businesses and customers.

This moment reminds me of the early days of the internet or the mobile revolution - transformative technologies that changed how business operates. But this time, the pace of change is faster, and the potential impact is even greater.

The companies that thrive won't necessarily be the ones with the most resources, but those that can learn and adapt the quickest. Success will come from understanding the new capabilities available, identifying where they can create real value, moving quickly but thoughtfully to capture opportunities, and building defenses against potential disruption.