Well, that would accommodate the biological reproductive differences between men and women fairly well (women‘s biological investment in pregnancy/child bearing >>>> men’s, while men can impregnate multiple women throughout one day vs becoming pregnant roughly once a year - hence women gravitate for the best genetical option available that provides the most security for her and her offsprings)
Which promotes harem structures and social unrest 😅 and yet, throughout most of our history most mean didn’t procreate while most women did - that also does indicate that some sort of this behaviour is part of our nature (yet being obviously perverted due to social media and dating apps giving people the illusion of endless possibilities and access to thousands/millions of people)
@kimmonismus Does anyone yet got a good understanding (apart from seemingly redundant benchmarks, who don’t seem useful anymore to assess model intelligence and usefulness) of the real world usability differences between fable 5.1 and Astra?
It’s not necessarily “just a mind moving faster than words can follow”. It just might be one way to explain as to why you can’t automatically infer low intelligence with someone who isn’t able to clearly and concisely express himself. But that doesn’t mean that everyone who’s stumbling over their own words is thinking in lightspeed
@DanDr1s But does anyone has experience with meta’s and google’s new models in real workflows? Cause benchmarks can’t really be trusted, since these optimized models all seem to show extremely positive results, yet sometimes do let down a lot (see the opus 5 debacle and meta’s llama 4)
@kimmonismus Well, fable 5.1 gains are on RL, Astra will be a completely new pre-train base model. Will be much more interesting to see what it’s capable of
@honorablesaint@bigswigger325 Well it does depend on which humans and at what time when taking about selection pressures for humans. There’s also forest hunting/stealth hunting that selects for different attributes vs long distance hunting in the open Savannah
Pretty much everyone thought so, even if it didn’t feel like 20x higher rate limits per week. I remember that I was quite disappointed trying it with fable for the first time since it didn’t feel like the usage jump that was promised and yet I trusted that it most likely (against my intuition) probably still is the case
Evolutionairy signals like the ability to provide and protect are typically perceived more attractive by females for obvious reasons, while men look out for market of fertility and the ability to raise children, which obviously makes sense as well. Why should men and women on average be attracted to the same signals? If that would be the case men and women would look, think and behave the same, but they are specialized for different tasks and show relevant mean differences in a big variety of traits. So how much money a woman makes and/or her career usually is less important to a man than it is to a woman