Background: Some data science, natural sciences, programming.
Familiarity/experience with Safety/ Alignment research: none, except for reading a lot about it.
I expected this to be a train the trainer seminar that starts at zero.
Perhaps it also was about me not pre-commiting to do 11 more months of more than a full-time job. Or not adding a CV, I did not cite any references when I have no clue which systems would handle my data and intelligence agencies would read it.
genetically-engineered grandchildren
What? We will have problems feeding the children, why would being genetically engineered have any advantage when real-world problems are not being solved, and instead are likely to grow for most of humanity, due to mismanagement, climate change, greed and xenophobia?
For individuals longevity and immortality are great. If you can live without your time, friends and family and cope with that.
For civilizations and in terms of fairness, this is probably terrible.
Already applied and was not considered as a good fit.
So not enough quality people made an attempt, and I am seemingly not quality people.
“in a way, o3 is more intelligent than any human” — seriously doubt this. It is more like Donald Trump, it knows nothing but tells you that everything is great and here is an inconcrete answer to your question, and now I ask you a follow-up question.
In nearly three years of asking ChatGPT, I leave the chat furiously in 80% of cases with a “fuck you, I am talking to a chinese room, not an entity that can tell me more than rubber duck debugging can”.
Fucks in the conversation start frequently at my second question, and make the chatbot reply more concise and sometimes even backtrack and take a different path, so I like to swear a lot.
For most parents, learning to cook and eat healthy should have stronger impacts than these genetic methods.
At 1% the cost.