Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G This is only a symptom of the problem. The end state is everyone being replaced … ytc_UgxsTYXwu… G I’ve worked in tech for decades, and I’ve seen both sides: innovation can libera… ytc_UgwubG4uB… G Zionist technology and trump about to approve a contract for Palentir to spy on … ytc_UgzDMhdGi… G I had never heard of greg rutkowski or sam does art before ai art. Its kinda lik… ytc_Ugx_1QblW… G why would anyone ask someone about AI who told engineers to print code on paper?… ytc_UgyLKrw_K… G Show Me How... I'm going to slightly adjust what I say to people when they ask a… ytc_UgwfHF7sG… G Facial recognition on the timeclocks at my job get me mixed up with others and o… ytc_UgxPqGdAa… G Although this was a fun conversation to listen to, I 100% believe that AI is not… ytc_Ugz-7Ymtf…
Comment
A lot of what she asserts is well supported, by generally, well over 99% of the energy/carbon costs are during training not during inference (end user usage). This assumes a baseline of google query vs GPT 4.x LLM query. We do need mire efficient ways to train AI, but AI has already come up with a new algorithm that dramatically reduces the cost/carbon impact of the highest non-AI use case: search/sort routines. Also, the current 10-fold increase in price/performance per year, coupled with the fact that “we have no moat, and either does Open AI”, means soon any kid in their bedroom, bad actor, small team from an adversarial country can engineer some thing far beyond today’s AI for 1/thousandth to 1/millionth the cost. What she has completely missed is that unlike Moores Law that takes about a lifetime to be over a million times more capable at the same price, AI will grow to that extent a couple times in about a decade. So, we are not “all in it together and able to decide together where it goes”, as individuals with a few thousand dollars will be able to do things far beyond what a team with a billion could last year. And maybe more importantly, a very advanced future version of where we are at is only a few years away given the double exponential growth of AI.
youtube AI Responsibility 2024-02-12T14:2…
Coding Result
DimensionValue
Responsibilitynone
Reasoningconsequentialist
Policynone
Emotionapproval
Coded at2026-04-27T06:24:53.388235
Raw LLM Response
[ {"id":"ytc_UgyhX1bLYVaXaWys16B4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"none","emotion":"outrage"}, {"id":"ytc_UgwJ6XXnt3BknYD75194AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_Ugyd0VOOFhIgKWV_qDN4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}, {"id":"ytc_UgyzuHjd9BKtUxlQSLt4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"ytc_UgzW5jSwYFEbumylX3V4AaABAg","responsibility":"ai_itself","reasoning":"deontological","policy":"liability","emotion":"fear"}, {"id":"ytc_Ugx2gp957etl9p3Ck1N4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"none","emotion":"resignation"}, {"id":"ytc_Ugw7HpySi8YMZLCCjNx4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}, {"id":"ytc_UgxMS6s7X58GmHFoiXl4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_Ugyx-4wX03RPyG2pmFN4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"}, {"id":"ytc_UgxzYJCJ70faZQaS5nF4AaABAg","responsibility":"company","reasoning":"deontological","policy":"none","emotion":"outrage"} ]