I wish there was more English premier league content here.
Posts by Recsysml
Posted here as well. www.linkedin.com/posts/gaurav...
I hope this gets the visibility to the the right audience
I can buy into that!
@bsky.app this is an example where if you want to provide a delightful user experience showing more content on swipe up either related to @petebuttigieg.bsky.social or the topic would make a lot of sense
@bsky.app when taps on the video an immersive viewer opens up. I recommend providing swipe up experience to show related videos preferably from the same author if available
@bsky.app great job on finally having a swipe up experience for more videos on immersive viewer
Can we try to understand why these people did what they did and understand their point of view?
Wow a masterclass in public speaking
How should we change our social media behavior? I remember how privacy and anonymity were primary user requests on forums like #discord for instance
An openness to experiment: contrary to popular perception bigtech’s money earning machines have not shifted over to LLMs but there is a top to down increase in the willingness to experiment new things. There is a lot more situational innovation in bigtech now.
Where can one try it?
Social media has a very high first Eigen value :)
@bsky.app please improve your SEO. @sungkim.bsky.social posted the same on threads and that is on google already if you search “zen c”
cc: @iwsmith.bsky.social
2026 will be the year where fluff is called out on Wall Street
Where can one track your administration’s success against the goals so far?
Reminds me of the line from Batman … heroes come in many forms youtu.be/DRMu1LA1gq4?... we are eternally grateful to you chairman Powell
Your campaign was built around listening to people, aggregating their voice and channeling those into your mandate. I am wondering how you will continue to listen to people as mayor and in particular how you plan to use technology like LLMs to scale the endeavor. 🙏🏾
A big change is coming in recsys. Long ago ranking models were actually ranking when it was doing pairwise ranking. Then it became an estimator model in the era of VM. Can we unlock all the RL advances of LLMs by making ranking a reward maximizing model again.
open.substack.com/pub/recsysml...
At some point it is just hubris that they know better.
#LLM vs #RecommenderSystems (Part 2/3) 👉 recsysml.substack.com/p/recsys-ret... is about Retrieval.
And this is where the analogy with LLMs becomes surprisingly tight.
Core idea:
Retrieval in recommender systems plays the same role as pretraining in LLMs.
We show clustering > "Semantic IDs"
Just to play the devils advocate, in tech this used to be the advice but recently tech leadership culture has changed to the 996 style. Do you think that is temporary or a precursor of a norm?
Wow 😮
So much work goes into making a clean ablation post like this. Admire your service
Fantastic progress!
Just so cruel what Russia is doing
Well played?
I don’t know much about kernels so forgive me for the noob question but what makes these kernels faster?
In this three part series I compare #LLM and #RecommenderSystems to show the gaps and opportunities. They are surprisingly fewer than one would think.
Part 1 open.substack.com/pub/recsysml...
Misdirection? So ppl talk about this and not real stuff?
Appeasing the king
Interesting that FBI report was leaked. I suspect they know how to not leak stuff