Big Creek Briefing

AI Solving Math Problems

Welcome to this week's Big Creek Briefing.

This week, AI crossed a line that most people weren't paying attention to. A Google DeepMind model solved an open mathematical problem that humans have wrestled with for decades. Meanwhile, OpenAI dropped a major ChatGPT upgrade, a popular developer tool got caught in a transparency controversy, and the data on small business AI adoption hit a new high. Lots to get into.

Let's jump in!

Big Stories of the Week

OpenAI Upgrades ChatGPT and Plans to Nearly Double Its Team: OpenAI released GPT-5.3 Instant as the new default model for ChatGPT this week, optimized for smoother responses and fewer unnecessary refusals. Internal evaluations show a 26.8% drop in hallucinations when the model is paired with web search. On top of that, OpenAI announced plans to grow its workforce from roughly 4,500 to 8,000 employees by the end of 2026, with the push focused heavily on enterprise sales and product delivery. The company is clearly moving from research lab to full-scale commercial operation.
Google DeepMind's AI Just Solved an Open Math Conjecture: AlphaEvolve, a Gemini-powered coding agent from Google DeepMind, did something this week that no human had managed to do: it discovered new mathematical structures that improved on long-standing open problems in complexity theory. To be clear, this wasn't AI assisting a researcher. This was AI advancing the frontier of human knowledge on its own. The implications for scientific research, drug discovery, engineering, and business analytics are significant and are only beginning to be understood.
Cursor's Transparency Problem: Cursor, one of the most widely used AI coding assistants among developers and technical founders, released its new Composer 2 model this week. The problem: the model was built on Kimi K2.5, a Chinese open-source model from Moonshot AI, and Cursor's leadership did not disclose that upfront. After being pressed publicly, executives acknowledged that roughly 25% of the model's foundation came from Kimi K2.5. For founders and business owners relying on AI tools, this story is a reminder to ask harder questions about what is actually powering the tools you depend on.

Main Insights

The Small Business Case for AI is No Longer Theory

The ROI case for AI in small business is no longer theoretical. A new survey found that 91% of SMBs actively using AI report revenue increases, and 68% of U.S. small businesses now use AI regularly, up sharply from 48% in mid-2024. That kind of adoption acceleration in under two years is almost unheard of for any business technology. The businesses still sitting on the sidelines aren't just missing out on efficiency gains. They're watching a growing portion of their competition pull ahead in ways that are becoming harder to close.

Founder Insights

Here's a stat that says something important about where we are culturally: half of U.S. small business owners report that the rise of AI inspired them to consider entrepreneurship as a path they hadn't seriously thought about before. That's not a productivity story. That's an access story. AI has lowered the barrier to starting and running a business so significantly that people who never saw themselves as founders are now launching companies. If you're already running something, that means your competitive landscape is growing. But it also means the tools making it possible for them are available to you too.

Quick Hit: One Prompt to Simplify Your Week

"You are a business strategist. Based on the following description of my business and my three biggest current challenges, identify the single highest-leverage AI use case I should implement first and explain exactly how to get started: [describe your business and challenges]."

AI Tool Spotlight: Missive

If your business communication is spread across multiple inboxes, platforms, and threads, Missive is the tool that pulls it all into one place. It's a shared team inbox that combines email, internal chat, and task management into a single workspace. Every conversation, whether it's a customer email, a team discussion, or a support ticket, lives in the same place and can be handled collaboratively without forwarding things around or losing track of who responded to what.

What makes Missive particularly relevant right now is its AI layer. You can generate draft replies instantly, get summaries of long conversation threads in seconds, translate messages on the fly, and build automated rules that route, tag, and respond to incoming messages without any manual work. For small business owners who are the first responder for customer inquiries, sales questions, and internal communication all at once, that automation layer is where the real time savings show up.

Missive is best suited for teams of two or more sharing a customer-facing inbox, and it's especially popular with agencies, consultancies, and service-based businesses. It holds strong ratings on both G2 and Capterra heading into 2026. You can explore it at missiveapp.com.

What We're Watching

Apple WWDC 2026 is set for June 8-12, with a heavy AI software focus

After a slower rollout than competitors, Apple is expected to use the conference to close the perception gap in generative AI across its entire platform.

AI model pricing is collapsing

Google's new Gemini 3.1 Flash-Lite is priced at just $0.25 per million input tokens, 2.5x faster than its predecessor. The cost of building and running AI-powered products is dropping fast.

AI solving scientific problems without human direction

AlphaEvolve's math breakthrough this week is not an isolated event. Expect to see more announcements about AI advancing research in biology, chemistry, and engineering in the months ahead.

AI startup valuations are commanding a 42% premium over non-AI startups at the seed stage

If you are building something with AI at the core, investors are paying attention in a way they weren't even 18 months ago.

The Wrap(TLDR)

ChatGPT got smarter and OpenAI is scaling fast. Google DeepMind's AI solved a math problem humans couldn't crack. And Cursor got caught not being upfront about what's powering one of its core features, which is a useful reminder to ask better questions about the tools you use. The adoption numbers are overwhelming at this point: 91% of SMBs using AI are seeing revenue increases. Half of new founders say AI is what made them consider starting a business in the first place. The tools are here, the data is clear, and the window to get ahead is still open.

If you want help designing AI workflows or understanding how these new models fit into your business, Big Creek Growth can help. Reply to this email and we will walk through it together.