raghavbali

Amazing work there @jrmyphlmn . Very recently I was thinking about how to preserve photographs I have collected over last 10 years in my external drive and instagram and unchecked SSD cards. Time to weave them all together. Bonus for me to be the 100th github starer on your repo

> With RSS, you subscribe directly to websites, blogs, or news outlets, meaning there is no middleman algorithm deciding what you see.

This enters a failure mode very soon, especially because most people using RSS-like technologies would typically subscribe to more sources than they can typically read through. Like it or not, _the algorithm_ does serve the purpose in prioritizing and discovery. The trouble, IMO, is with the objectives for these recommendation and ranking algorithms.

A middleman/aggregator who is paid by subscribers would be incentivized for the users, a marketplace-like aggregator would always have trade-offs.

> Can’t convince someone of the former when they don’t even understand that the computer is the box attached to the monitor, not the monitor itself.

This is so true and more often than it should be. I am going to use this phrase !

Can't express it more clearly than this. Data structures are just one part of the story not the only spot where the rubber meets the road IMO too. But going back to top of the thread, for new projects it is indeed steps 2 and 3 that consume most time not step 3

Coool. I remember when the OG pebble launched but I couldn't get one for myself (it wasn't available in my region and my pocket money didn't allow for it either ;) ). Looking forward to this #bitesNailsFuriously

> Unfortunately if you naively quantize all layers to 1.58bit, you will get infinite repetitions in seed 3407: “Colours with dark Colours with dark Colours with dark Colours with dark Colours with dark” or in seed 3408: “Set up the Pygame's Pygame display with a Pygame's Pygame's Pygame's Pygame's Pygame's Pygame's Pygame's Pygame's Pygame's”.

This is really interesting insight (although other works cover this as well). I am particularly amused by the process by which the authors of this blog post arrived at these particular seeds. Good work nonetheless!

Maybe I missed something but this is a round about way of doing things where an embedding + ML classifier would have done the job. We don't have to use an LLM just because it can be used IMO

Nicely summarised. Another important thing that clearly standsout (not to undermine the efforts and work gone into this) is the fact that more and more we are now seeing larger and more complex building blocks emerging (first it was embedding models then encoder decoder layers and now whole models are being duck-taped for even powerful pipelines). AI/DL ecosystem is growing on a nice trajectory.

Though I wonder if 10 years down the line folks wouldn't even care about underlying model details (no more than a current day web-developer needs to know about network packets).

PS: Not great examples, but I hope you get the idea ;)

why not fix the calculator in a way that avoids/mitigates scenarios where users get to wrong quotes and then do an A/B test? This setup seemingly tilts towards some sort of a dark pattern IMO

Yet to go through in detail but this is really powerful. Initiatives such as these are what we need to further democratize DL. Kudos team

the browser wars are heating up again! nice

This is such an amazing way to teach so many of engineering topics. Time to dust off that mechanix kit!

yeah packaging the data and updating it async does make a lot more sense. Also, I guess its fine that it doesn't know it all but covers a good percentage of requests

Thank you for the nohello.net thing, I am usually pretty awkward when it comes to starting conversations and but I guess I never paid attention to why that was the case. The discussion on this thread clears out the impression I had that it is usually rude to directly jumping to the question/task. I got my queue! :)

+1 and also the fact that Google Pixel's "Now Playing" feature is such an amazing application of the same idea. Though I wonder how different are their implementations

this is sweet... thanks for the suggestion

Could you elaborate that a bit more on this? I am on SDE side and keen on understanding the PM motivations as well

I like the part " _microservices is basically SOA_ " , want to say that loud so many times myself :)

vector search is everywhere nowadays. Glad to see another success story. Kudos to the team

that is some really cool work there. Thanks for sharing

Kudos to @sjdz for this. Really like the aesthetics. The ability to change formatting elements for ease of reading is bang on.

you just nailed it with the power law example. But how would you explain the _rags-to-riches_ scenario?

My recent book explores different generative models and their applications. Bonus : Code is available on github : https://github.com/PacktPublishing/Hands-On-Generative-AI-wi...

Kudos. This is quite impressive @skzv

Thanks OP for sharing this interesting work/stack. My kindle has been pretty useless (I prefer reading novels on paperbacks), plus I have been itching to try our obsidian for quite some time as my pile of bookmarks is becoming quite difficult to manage too. Will give it try (or a derivative of it)

Wow, the journey is inspiring. thanks for sharing.

Something I have been looking for quite some time. Thank you for this.

There was a recent Hindi/Bollywood movie on the same subject showcasing the plight of the living dead. Kaagaz (Hindi for Paper): https://en.wikipedia.org/wiki/Kaagaz

Well, I agree to your point that the explainability is only as good as the features used in the model. Yet it is important to attribute the correct set of features (along with quantification) for a given instance. The SHAP framework does that to a good extent but it's focus is not towards helping with identification of model parameter related issues. This work here seems to focus on the later and hence I found it a bit different and a refreshing perspective towards the explainability aspect.

Kudos to authors for such detailed work. I am yet to go through in detail but the way they have approached the problem certainly is refreshing. Any thoughts on how to extend this to non-vision architectures?