As AI model performance converges, the key differentiator will become memory. The accumulated context and personal data a model has on a user creates a high switching cost, making it too painful to move to a competitor even for temporarily superior features.
A new, specialized role will emerge within large companies, combining functional expertise (e.g., HR, legal) with "vibe coding" skills. These individuals will act as internal consultants, building bespoke AI applications directly for departments, bypassing traditional IT backlogs.
Nimble small and medium-sized businesses will increasingly use AI to build custom internal tools, especially for CRM. They will opt to create the 20% of features they actually need, rather than pay for complex, expensive enterprise software where they ignore 80% of the functionality.
Major AI labs will abandon monolithic, highly anticipated model releases for a continuous stream of smaller, iterative updates. This de-risks launches and manages public expectations, a lesson learned from the negative sentiment around GPT-5's single, high-stakes release.
Individuals will use AI to build bespoke software for personal use. A subset of these tools will find a niche market, creating entrepreneurs who operate outside the VC-funded, subscription-SaaS model, potentially favoring one-time purchase models due to low development costs.
The most significant gains from AI will not come from automating existing human tasks. Instead, value is unlocked by allowing AI agents to develop entirely new, non-human processes to achieve goals. This requires a shift from process mapping to goal-oriented process invention.