Experts are criticizing the excessive concentration of tech companies on achieving AGI, noting that ignoring structural barriers—regulations, economic incentives, and the complexity of physical systems—could slow down the real-world implementation of AI in key industries.
What Happened
An article by The San Francisco Standard examines the gap between progress in neural networks and their integration into the real economy. Specifically, it notes that in biopharmaceuticals, despite AI advancements, drug development remains extremely expensive (exceeding $1 billion) and lengthy (about 7 years), and the existing patent system incentivizes only the optimization of known molecules rather than the search for fundamentally new biological targets.
Context
Silicon Valley is focused on creating Artificial General Intelligence (AGI), but a technological breakthrough in model "intelligence" is only part of the solution for capital-intensive and heavily regulated sectors like medicine or construction. Current economic models and patent laws create an environment where AI companies tend toward "herd behavior," choosing less risky optimization tasks instead of deep scientific research.
Why It Matters for the Industry
For the industry, this means a necessary shift from creating "pure intelligence" to developing Vertical AI—specialized models adapted to specific regulatory and scientific requirements. There is a risk of unjustified investments in general models that do not account for the specific workflows and legal constraints of the real sector.
Why It Matters for Users
Readers should understand that progress in medicine or industry will move slower than predicted by AGI-oriented optimists. A focus on powerful models without considering infrastructural and legal barriers could lead to the emergence of many "thin wrappers" over neural networks that merely simulate solving complex tasks without changing their fundamental cost.
Sources
Author
Look at AI, Editorial Staff