- AI based on pattern matching is a limited form of lower animal intelligence that has little to do with human thinking
- AI will never live up to the hype because the problems it hopes to solve are mostly not valuable needs of humanity
- There is no such thing as “unbiased” AI. Claims about creating “unbiased AI” are simply disguised attempts at backdooring a particular political view into industry.
- AI researchers have demonstrated no ability to forecast which problems that are not solved are almost solved. Until something is demonstrated to work, assume it might not work for decades or never work at all.
- Investment in tech or AI is not based on what ideas will make a major impact on the economy. Investments are made based on what investors suspect can be sold to the public as an exciting prospect in 3-4 years.
- Most uses for AI are relegated to entertainment. Partly this is because the biggest bottlenecks of real economic growth are not planning and decision making but rather the cost of energy. Entertainment is a continual zero-sum game which can absorb an unlimited amount of resources.
- The biggest value of generative AI is as an information retrieval tool. And therefore the best uses are in education and in providing simulated and downgraded, specialized human services to people who currently cannot afford to pay for those human services (e.g. legal advice).
- Innovations in information technology are largely solutions seeking problems rather than the other way around.
- Most AI inventions are limited and brittle toys. The necessary work to build products around them are often too difficult and expensive to make using them practical. The innovations we need most are in (the less exciting) product design and AI integration patterns, not new AI breakthroughs.
- The biggest technology advances tend to come from unexpected and ignored areas such as material sciences. “One Word: Plastics.”
- In the book Dune, the writer Frank Herbert cleverly solved the problem of not having to predict the future of computer technology by introducing the Butlerian Jihad, a mass uprising against computers and the subsequent adherence to the idea: Thou shalt not make a machine in the likeness of a human mind. Think on that.
- There is an amazing and unforgivable lack of focus on creating tools that actually raise productivity in a measurable way.
- Raising ultimate economic productivity is impossible if technology is applied to jobs that create no real economic value to begin with. Attention often should be directed not at doing some tasks faster but rather not doing unnecessary tasks.
- There are many different areas of AI. We are currently making rapid progress in the pattern recognition area and mostly ignoring the others. That will someday change.
- AI utilizing natural language is somewhat of an illusion. What it exposed is that language itself is not really a manifestation of intelligence. Language is just something that has little use to a being that lacks intelligence and so is correlated with it. Intelligence is below the level of language.
- The AI field has learned that they get more attention and investment from the populace when they anthropomorphize AI. The public’s understanding of AI will always be driven by Hollywood and researchers must therefore fit their ideas into that mold. Otherwise they are just doing boring computer science.
- Five years from now the most impact from AI will be in? embarrassing areas such as pornography, virtual “companions” and in zero-sum areas such as cyber-crime and cyber security.
- In 10 years, the biggest impact will be with military robots. Human soldiers will largely be replaced by robotic soldiers of some type. Wars of the future will be less bloody affairs and will look more like Pokemon or BattleBots and that won’t be a bad thing.
- Generative AI is very good at generating written content but mostly content that no one really needs to read anyway. Better uses of generative AI will be in reducing the amount of content that we need to consume.
- Generative AI won’t create value by thinking for us but rather helping us think better and stay focused on areas that require our attention and less focused on distraction.
- For the most part, AI is not “coming for our jobs” much faster than most technology advances. This is especially true of good jobs. Partly this is because making transitions of any kind takes investment which is expensive and can only happen so fast.?
Chief Growth Officer helping industries become more profitably sustainable thru proven digital transformation solutions with #embeddedAI
1 年always insightful David
Senior Director Mondelēz
1 年This is very good, Dave. Curious to see what kind of reaction it would get in certain forums.