Ertan Dogrultan (@ertan) on Farcaster

Ertan Dogrultan

@ertan

Platform Engineering at Replit. I like systems.

San Mateo, CA, USA

172Following 615Followers

Casts

Casts + Replies

ertan

has anything in human history been researched as intensely as AI?

ertan

I wish I talked about my dislike of Taylor Swift more before it became a partisan issue

ertan

farcaster frames and replit agent has a really interesting convergence point towards ephemeral and/or JIT apps.

ertan

9/9 was my 15th anniversary of moving to the US for good. thankful for all these years, opportunities. glad to have a wonderful family with 2 kids and work on some of the most interesting problems. I'll forever be thankful and I hope my kids don't take it for granted.

ertan

really excited about this launch we've been working for a while. truly feels like we are at the cutting edge of what's possible in programming

Amjad Masad

@amasad

AI is incredible at writing code. But that’s not enough to create software. You need to set up a dev environment, install packages, configure DB, and, if lucky, deploy. It’s time to automate all this. Announcing Replit Agent in early access—available today for subscribers:

ertan

party like it's 999

teenage.engineering

EP–1320 medieval

ertan

my favorite thing is to cancel Audible every 3 months and get a deal like this within 1-2 months

ertan

the feedback loop for getting a power user badge requiring interactions from others seems to be broken. my casts aren't visible because I don't have a badge even though the content quality is decent. it's skewed heavily towards power users with high follower count and I worry the algorithmic feed will not solve this un...

ertan

this is incredible!

Ahmad Al-Dahle

@Ahmad_Al_Dahle

With Llama 3.1, we’ve updated our license to allow developers to use the new models (including 405B!) to distill smaller, more efficient models that can be fine-tuned for specific applications. We’re expecting this to ignite new innovation from all corners of the community.

ertan

when AI video generation matures and gets cheaper, one should feed it threads of different sagas from X to generate series on the fly and disrupt all streaming services. it's mental already. I can't imagine watching this shit in video form.

ertan

dev

what is the best incident review process you participated in? what was the format and why was it good? I'm looking to improve our retro process so looking for ideas.

ertan

good high level overview of encoder, encoder-decoder and decoder only models

yitay.net

What happened to BERT & T5? On Transformer Encoders, PrefixLM and Denoising Objectives — Yi Tay

ertan

I went back to X to stay up to date with recent AI literature and catch up for the last few months. now that the election season is coming, things have gotten so unhinged that I've decided to come back to Farcaster and be the change I want to see. will try to post more often about the topics I'm interested in.

ertan

Ertan-kun is a baka dev who can't even find the right button without an LLM, uwu~ Maybe I should make an AI to tie my shoelaces too! Create a Waifu based off your FC profile with WaifuCast!

waifucast.novadao.io

WaifuCast

ertan

RAG is the practical way of making inference unit economics work for many applications

Rulin Shao

@RulinShao

🔥We release the first open-source 1.4T-token RAG datastore and present a scaling study for RAG on perplexity and downstream tasks! We show LM+RAG scales better than LM alone, with better performance for the same training compute (pretraining+indexing) https://t.co/5QChH9Jn6u 🧵