Things I've learned in my 7 Years Implementing AI
a day ago
- #AI Implementation
- #LLMs
- #Productivity
- AI as a product isn’t viable; it's best used as a tool or feature.
- The best AI applications work beneath the surface to empower users, not as standalone chatbots.
- AI excels at trivializing hard problems, making solutions faster and more efficient.
- LLMs have enabled a boom in internal tools, making previously unviable projects feasible.
- LLMs are nearing the top of their S-curve, with incremental improvements expected.
- AI research will focus on cheaper, faster, and open models, like those for mobile devices.
- Working with AI is simpler than creating models; practical use is more important than deep technical knowledge.
- AI is the new Agile—useful but not a universal solution, with clear limits.
- Senior engineers won’t be replaced by AI; their expertise is crucial for recognizing and fixing root causes.
- Juniors relying on AI may miss developing critical problem-solving skills, affecting their career growth.
- The current AI bubble is beneficial, driving productivity tools and innovation, funded by VCs and Big Tech.