W01 - Reviewing Changes in the AI Field in 2024

Looking back at changes in the AI field in 2024 from my perspective, I focus only on a few leading companies and limit the scope to information distribution and coding.

First, OpenAI. Major turnover in the core team exposed governance challenges; the returns on the bet on pretraining are shrinking, and a sustainable business model has not yet been established. The company repeatedly missed opportunities in niche markets: search was overtaken by Perplexity, and coding was overtaken by Claude. Even so, ChatGPT’s monthly active users are approaching 600 million, far higher than the runner-up. This shows that in the consumer market, brand recognition drives adoption much more than technical differences—most users can’t tell which model is better, so choosing an AI product by brand remains the most efficient approach.

Claude demonstrated a clear strategic play in 2024. Its foundation model focuses on coding, the leadership is stable, and it has Amazon’s backing. After Sonnet 3.5 was released, Claude gained an absolute advantage among developers, and more developers means more model API calls. With the launch of the MCP protocol, its developer-centric growth strategy became even clearer. This approach resembles Stripe’s: capture developer mindshare first, then leverage that to expand the broader commercial ecosystem.

Cursor and Devin emerged as dark horses in the coding space, underscoring that context is key. I haven’t tried Devin; it’s primarily B2B and reportedly feels more forward-looking. Cursor’s superior developer experience compared with Copilot comes down to how it handles and interacts with context. Of course, in some respects it still falls short of humans. Its understanding of temporal and spatial context is limited—for example, when the same line of code undergoes different changes, the time context is an important analytic factor; lacking that makes it hard to accurately infer relationships between edits. Its grasp of multiple background relations and long-range reasoning—such as the composition of a monorepo, which package manager is used, or which in-house infrastructure libraries are introduced—are all everyday developer knowledge that Cursor cannot reliably understand yet.

Software could be the largest source of incremental value in the next era. Practitioners can feel that software production models and costs are on the brink of a major shift. The mobile internet produced content; the next era will produce software. Increasingly, people without CS backgrounds will choose to develop software themselves to solve problems. Just as traditional media gave way to individual creators where anyone can publish, traditional developers will be followed by self-developers where anyone can write software.