In psychology, cognitive dissonance is the discomfort from holding two or more contradictory thoughts. The term describes AI today. To leverage AI and thrive in our AI journeys, we need to live with the discomfort that comes with understanding of the strengths and weaknesses of AI.
ChatGPT, Gemini, Claude:
- Thought – AI improves coding, writing, and summarizing content productivity
- Thought – AI hallucinates including 30 percent to 90 percent on references and struggles with context
Chatbots for advice:
- Thought – Nearly 75% of study participants turn to AI for advice, 39% described AI as a consistent and dependable emotional presence
- Thought – Your favorite chatbot is lying all the time
Large Language Models:
- Thought – A.I. is getting more powerful
- Thought – but its hallucinations are getting worse
AI Agents:
- Thought – AI agents are the next big thing and have moved from Chatbots to taking on enterprise workflows.
- Thought – AI agents not ready to solve a savvy Tech podcasters problems and risk selling a Chevy Tahoe for one dollar.
AI Reasoning Models:
- Thought – AI achieves 89th percentile on competitive coding, top in USA Math Olympiad (AIME), and exceeds human PhD-level with physics, biology, and chemistry problems (GPQA).
- Thought – Apple study found large reasoning models (LRMs) face a complete accuracy collapse beyond certain complexities.
Autonomous Vehicles:
- Thought – In 2016, a self-driving Budweiser beer truck makes a delivery 120-miles away. Waymo provides 250,000 autonomous driving taxi rides per month across four cities.
- Thought – No full autonomy (SAE Level 5) vehicles yet. A study found Waymo charges about $10 more per ride than Uber and Lyft which doesn’t sound profitable yet.
Photo Credit: Author generated with ChatGPT. AI image generation is amazing, though it can be a struggle to get precisely what is wanted.
Recent Comments