@dwainmcfarland@jagibbs_23 You think it’s ok to fade 1st team camp target distribution? I know Gainwells historical TPRR is nowhere near Warrens. It isn’t sitting well in my gut knowing Arthur’s past.
https://t.co/KErijy2qIM
TPRR stabilizes after 200 routes, YPRR after 300 routes IIFC so this is actually not a small sample. However his usage is the issue, it doesn't sound predictive. We saw this before with Skyy Moore, where Andy just gave him all the designed volume and boosted his YPRR but Burden is better than Skyy.
The only similarity between ai and covid is that this moment in ai feels similar to the moment in covid when forecasters (like @balajis) first figured out that covid was going to be massive and the public was unaware/didn’t believe. Most don’t know coding has been effectively automated and that every other knowledge job is about to get automated as well.
As of today, only two Canadian teams occupy playoff spots. So which one leads all Canadian teams in our monthly @Sportlogiq NHL Power Ranking?
The Montreal Canadiens? Nope.
The Edmonton Oilers? Guess again.
Why the Ottawa Senators sit in the top spot: https://t.co/SuqZya0QZk
I'm building The Record to solve this problem. Here's how:
- We use an LLM council + wisdom of the crowds to analyze the truth of individual claims (with scores weighted by model/users historical track record of predicting truth)
- When a source has >25 claims, we aggregate their scores to create a "Trust Battery" which gives you a real quantifiable metric for how reputable they are
- Use this dataset to create and sell media products for consumers (newsletters, fact checking tools) and businesses (data for ai labs)
We've scored some of mainstream media so far and have seen some interesting results - wapo appears to be pulling away from the crowd in terms of reputability. Would love to get your feedback.
We "fact checked" them with our transparent truth scoring system (LLM council + wisdom of the crowds) and they're sitting at 78% True overall. Not good enough. Wapo is looking better at 86% True overall.
However neither compares to @TheRecordApp, where every claim made is >90% True.
Media is a large highly fragmented industry w low NPS right? @TheRecordApp's vertically integrated solution will win. Would appreciate your feedback Keith!
This is a perfect example of why people do not trust the NYTimes anymore.
@TheRecordApp's transparent truth scoring system rated this claim 3% True and we have the NYT's overall trust battery at 78%.
That is not good enough to earn people's trust.
Laughed seeing NYTimes say Alexandr Wang "isn't an Engineer"
When we hired him as a 17 year old software engineer, he had ranked as the 10th best competitive programmer in the US at IOI
@MattHarmon_BYB I like it descriptively, less so for predictive value. Good way to answer the question of did you convert your opportunity into production. Probably not super predictive cause it depends on what opportunity you were given.
@ScottBarrettDFB We need a @FantasyPts WR score combining separation, yac and catch point. Could be the most predictive metric non production stat for YPRR.