[flagged]
AI "reasoning" is the model babbling to itself, using statistically likely sequences of words that appear in the training data. Of course that will often produce valid chains of logic, but not because it in any way understands the principle behind it. It is ONLY modelling LANGUAGE, not any of the mental processes that guide the use of language in humans (and which might even exist in animals not capable of language).
When asked to carefully check itself, it might spot errors because their structure matches common examples of fallacies, or it might not, and instead insert some further bullshit like "Assumption Verified — 100% correct! It all checks out [insert some emojis here]". That then becomes part of its context, and further boosts its unfounded confidence.
I joke that chat GPT is my child's primary pediatrician. It's not perfect but humans aren't either and it makes fewer mistakes than the Clinic and available in seconds 24 hours a day.
That's the difference between a baby that is 20% underweight and dying versus its birth weight.
I made them show me how they did it twice before leaving.
After I left I got a call from the doctor apologizing that all of their neonates weights were wrong that morning because their staff don't know the difference between decimals and ounces
Humans are incredibly error prone. Modern llms are far better than that.
When my wife was pregnant, the OB was asking an LLM for medication contraindcations while we were in the office. She showed us the model she used.
LLMs make mistakes, so do humans. Especially humans who dedicate <15min to a patient and often are reading the charts/info in front of the patient for the first time (during that short window).
It's frankly insane to say an LLM is your child's primary pediatrician.
You're looking forward to the value of human life under capitalism going to zero?
What do you think happens next, does capitalism go away or do humans?
That said, I think an awful lot of children today are going to grow up under the shadow of AI and they will not be grateful for it.
The future will be different. Hopefully the good ways will outweigh the bad.
Im fairly excited about AI integrated child education. I think it opens a lot of opportunities for kids to move and explore at their own pace- Hopefully with a caring human involved in some areas.
Maybe we can get the good ending!
I see that phones and tablets are absolutely cooking the brains of my nieces and nephews, I don't have any hope that adding AI to the mix will help that any
The downside is that this raises the developmental ceiling and lowers the floor, but those kids will have to find a way to coexist as adults.
Some parents will have purpose built llm tutors, while other parents will just cut their kids loose with grok and Twitter.