Arghya Sarkar

coordination tech! @orveltapp

[Arghya Sarkar @sarkarghya]

@saltyAom it wasnt random. xiaomi just released their version of mac mini

[Arghya Sarkar @sarkarghya]

@MilksandMatcha Got it so I will use ‘ —‘ or ‘— ‘ to make my em dashes look human

[Arghya Sarkar @sarkarghya]

@theo you don’t need to look far, just look at frontier math. One I think of are the hilbert problems. You can translate these problems into code (initial state), llms can give you a satisfying initial condition (example code) and ofc it’s verifiable!

[Arghya Sarkar @sarkarghya]

@samlambert Just try spacetimedb bro… 🤡

[Arghya Sarkar @sarkarghya]

@postimortem @mgostIH Back to if else statements. DAGs for you ummmah

[Arghya Sarkar @sarkarghya]

@MrinankSharma Good luck mrinank! I’m glad you could find the courage and resources to take this decision

[Arghya Sarkar @sarkarghya]

@wyatt_benno I’ll play with your repo when I have some time. Also how your work is different from Ezkl?

[Arghya Sarkar @sarkarghya]

@PlastiqSoldier @Mihonarium One of the most important developments leading up to agi (if it’s even possible) would be recursive self improvement. anthropic engineers a yearish ago agreed with me that learning from user data could achieve that albeit to lesser extent. Likely changed minds since then

[Mikhail Samin @Mihonarium]

Many in my community hold Anthropic in high regard. Sadly, they should not. I wrote a post showing why.

Anthropic in its current form is not trustworthy. The leadership is sometimes misleading and deceptive; they contradict themselves and lobby against regulations just like everyone else, while not really being accountable to anyone except perhaps their investors.

The post discloses a number of facts that had not previously been reported on and combines them with publicly available information in an attempt to paint an image of Anthropic more accurate than the picture Anthropic’s leadership likes to present.

Read: https://t.co/IjKE5VQZC0

[Arghya Sarkar @sarkarghya]

@Timur_Yessenov @_avichawla Ai bro discovers inverted index

[Arghya Sarkar @sarkarghya]

I am trying to figure out a way to make ephemeral human verification possible. Think world coin protocol but temporary

[NIK @ns123abc]

https://t.co/XR3flatMbW

[Arghya Sarkar @sarkarghya]

@ceo_clickhouse @mattturck @ClickHouseDB @langfuse Congrats! Please keep open sourcing your pipelines to iceberg and such

[Arghya Sarkar @sarkarghya]

Feverup is the app btw

[Arghya Sarkar @sarkarghya]

I hate vibecoded apps. Security aside can you imagine a single website using 3 gb?

[Arghya Sarkar @sarkarghya]

It is not money thats disproportionately distributed but the benefits of capitalism

[Arghya Sarkar @sarkarghya]

Singularity may look nothing like what industry leaders predict. The coming years will be pivotal in determining whether its trajectory benefits humanity as a whole or consolidates power for a select few corporations

[Leonard Tang @leonardtang_]

session #2 of the NYC AI Reading Group w/ @qw3rtman @willcb is in order!

> where: @haizelabs hq
> when: thursday 5/29 @ 7:30 pm
> what: sft memorizes, rl generalizes: a comparative study of foundation model post-training
> who: awesome people like yourself :^)

> also: PIZZA🍕