@TheWarningTake I’ve always had an issue with that experiment. Researchers started with just 4 pairs of mice. Lot of inbreeding that could confound all those results
But I agree in principle.
@wireheaded@ggerganov OP is showing the 27B param quantized Qwen. Frontiers are many times larger but even so, the 27B model outperforms Opus 4.8 here (estim. 1.5-2T params, 55.5x-74x larger).
Your own picture also shows the Qwen 3.8 2.8T param full model as third in the rankings.
Fantastic piece! What happened when a stubborn Jew refused to play along with sanctimonious clueless haters, exposing instead their shallowness and ignorance
"I went to an anti-Israel dinner party. It did not go as planned"
https://t.co/mJnMCtdG9A
@JoeA_NFL I see a path for Tyson if the Saints can plow through the NFC south.
Don’t know how any of the top three are the top three though. We saw how a stud rookie RB in a bottom 5 offense performed last year, Mendoza isn’t playing for at least half the szn, and Tate has a Ward problem
@RadCentrism@jakozloski@RealAdamB1 For sure, and need-based aid can still result in $10k’s out of pocket in cases, especially for those attending out of state. Versus a full / 75%ish ride elsewhere (or in-state) it’s often not a hard choice to make.
@RadCentrism@jakozloski@RealAdamB1 This would make filtering by school a noisier proxy, assuming your goal is to find people in similar IQ bands.
I concede my second point though, you’re right there.
@RadCentrism@jakozloski@RealAdamB1 Portraying the slightly above average student is more relevant than 1-2%. They freq. choose between paying much more at a selective or out-of-state school that offers them little merit aid and accepting a major scholarship from a less selective, “worse” college.
@RadCentrism@jakozloski@RealAdamB1 -Inability to get sufficient financial aid to go to a school that (stereotypically/heuristically) matches that persons SAT/ACT score. In turn, going to a “lesser” college for less $.
- Undergraduate student IQ has been suggested to vary ~1 StDev across fields of study.
@vibecodeguild@PatriciaMe39871@sama 4o was a model that was over-agreeable and lay on super thick emotional verbiage/“support” type language onto users. Many got addicted to its behavior and developed parasocial relationships with it.
I believe a neologism was offered for them somewhere: “sycofanatics”
@emanningwrites We probably only get 1.5 years with him before we have to try and flip a guy who’ll easily be asking for 13M+
If we don’t win with him while we have him the value in taking him becomes a net minus. I’m hoping more toward a Sharks NHL-ready prospect trade for Nikishin.
@emanningwrites Im actually fine watching the madness— pretty entertaining. Interestingly I’m becoming increasingly sure that I don’t want Werenski to the Canes. These AAVs are insane (Trouba?? WTAF) , + if he’s being as much of a diva as rumors are saying he is right now … 1/2