we've been tracking the deepseek threads extensively in LS. related reads:

- i consider the deepseek v3 paper required preread https://github.com/deepseek-ai/DeepSeek-V3

- R1 + Sonnet > R1 or O1 or R1+R1 or O1+Sonnet or any other combo https://aider.chat/2025/01/24/r1-sonnet.html

- independent repros: 1) https://hkust-nlp.notion.site/simplerl-reason 2) https://buttondown.com/ainews/archive/ainews-tinyzero-reprod... 3) https://x.com/ClementDelangue/status/1883154611348910181

- R1 distillations are going to hit us every few days - because it's ridiculously easy (<$400, <48hrs) to improve any base model with these chains of thought eg with Sky-T1 recipe (writeup https://buttondown.com/ainews/archive/ainews-bespoke-stratos... , 23min interview w team https://www.youtube.com/watch?v=jrf76uNs77k)

i probably have more resources but dont want to spam - seek out the latent space discord if you want the full stream i pulled these notes from

"President Walt Disney Pepsi Comcast has done wonders for the economy! Being that it now is...the economy."

One way that works moderately well is to mark certain topics as not interested when they are presented and like and watch the other topics for a longer time.

I've used the above to keep my YT landing page relatively clear of things I'm not interested in.

Once in a while it will suggest you a highly popular or interesting off topic item, but yea that is once in a while.

Basically what I'm saying is that we need to train their algorithm to our tastes through watching and marking things as not interested, don't show me this channel etc.

it's intentional impossible by their design. they need to mix content because they only have so much advertisers for your profile on each topic, so you must move around as they don't get paid for many displays.

but you shouldn't be using their clients anyway.

Given what Broadcom did to almost all of VMWare's products, I'd be a lot more worried about PCF than really anything else.

I've been writing an email newsletter on backend dev for the past few years - https://quastor.org/

Basically, I look through all the Big Tech engineering blog posts and look for interesting posts. Then I'll write up an article talking about the tech stack and the main takeaways.

I've been learning Adobe Illustrator + After Effects recently and am now starting to add some visualizations + graphics to accompany the articles.

I make more than $500 per month but this is more of a passion project and the income is honestly not competitive (at all) with what I earn as a dev.

It also gives me a way to learn about marketing/growth + monetization. Hard to learn that stuff w/o doing.

I started SupeChoiceForm as a passive income project many years ago: https://superchoiceform.com.au/

It started as just a content site, with google Adsense. Making about $1200 per month from ~30k visits.

It kept growing in popularity over about 2 years and I finally decided it might be worth some more effort.

So I turned it into a web app that took an employee through the process of gathering/generating all the paperwork they require to provide their employer with their superfund details. Effectively just digitising the paper form process with some smart automations.

(Superannuation is the 401k of Australia, but probably more widespread as it’s mandatory)

That was 2x MVP’s ago. Last year I got two trust partners / co-founders involved and we’ve turned it into an embedded service that can drop into any payroll/hr provider.

Raised $1M and now we’re making a lot more than $1200

All because I was nervous about getting Adsense banned for no reason. So thanks google!!