SDE • Prev Lead Collab @oppiaorg • @microsoft SAmbassador • Blogger (1.1M+ Reads), Mod @ThePracticalDev / @freeCodeCamp • AI/agent projects beyond the obvious
Kathmandu, Nepal
- Tweets 275
- Following 157
- Followers 110
- Likes 1,016
Pinned Tweet
Guys, do check it out ✌️
"Thinking differently" with Neovim by @shricodev #vim #linux #productivity
Full Post: dev.to/shricodev/neovim-make…
2
5
566
Shrijal Acharya retweeted
I sincerely believe Go-lang and Rust will be the language of the AGI.
46
18
490
27,464
one great way to test AI models locally!!
Sometimes you might want to run AI models locally without bloating your computer or taking up too much space.
And you can do this using Azure + Docker to set up a virtual machine that runs the models inside a container.
Here, @shricodev walks you through the process step by step.
freecodecamp.org/news/host-l…
1
36
Shrijal Acharya retweeted
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
20,620
167,418
802,959
173,934,399
Shrijal Acharya retweeted
GPT-6 Astra has done incredible things I never thought a model could do. It has also done some of the stupidest things I've ever seen a model do.
Generally speaking, Fable 5.1 just does what I ask.
408
132
5,034
732,401
If you’re working with Kubernetes, Keda and all sort of agentic AI workflows, this could be your read.✌️
You’ll learn to build durable autoscaling agent runtime.
Long-running AI agents need to be able to handle crashes and scale to fit workflows.
In this guide, @shricodev shows you how to build a durable, autoscaling agent runtime.
You'll use Temporal for retries and recovery, Composio for tool use, Docker and Kubernetes for deployment, and KEDA to scale workers based on workload.
freecodecamp.org/news/how-to…
74
Shrijal Acharya retweeted
Sometimes you might want to run AI models locally without bloating your computer or taking up too much space.
And you can do this using Azure + Docker to set up a virtual machine that runs the models inside a container.
Here, @shricodev walks you through the process step by step.
freecodecamp.org/news/host-l…
3
41
366
14,370
📰👨💻 I just published #Angular #TypeScript Trends #418 dormoshe.io/newsletters/418?… @shricodev @thepracticaldev @SamchonGithub @TomWebwalker @thegdsks @ErikCH @nyaomaru_dev @dormoshe #dormosheio #javascript #frontend #webdev
1
2
645
Wow! Got this for the first time after years of writing.
Thanks @ThePracticalDev 🫡
Congrats to our top 7 authors!
This week: building a LAN speaker, digging into what is causing the latest Big Tech layoffs (it's certainly not financial troubles), and much more.
Great job @shricodev, @ahmershah29, @tibtof, @minirobbo, and the rest.
dev.to/devteam/top-7-feature…
1
2
392
Shrijal Acharya retweeted
Nvim 0.13 gained vim.ui.img, a stdlib interface for displaying images.
- useful for markdown files, directory browse, picker preview, ...
- ~small interface + impl cost. mostly a protocol change.
- placement will be handled by extmarks or floating windows.
- no builtin codec, not going to vendor a jpeg library.
the vim.ui.img API for displaying images is now merged to @Neovim master.
the result is a nice simple interface, despite a lot of pull in the other direction.
idk why, but somehow these PRs for new modules always start with the most insanely complex implementations, instead of asking "what are the primitives that we absolutely need", and leaving the rest for later.
this PR started with:
- support for sixel, iterm2, kitty
- new "provider" layer
- async support??
- different concepts and "address spaces" for placement vs data
and it ended with:
-- Create:
local id = vim.ui.img.set(…)
-- Update:
vim.ui.img.set(id, { row = 8, col = 12 })
-- Get:
local opts = vim.ui.img.get(id)
-- Delete:
vim.ui.img.del(id)
yeah, an LLM would have been fine (ecstatic!!) with rev1. think about that.
5
25
275
26,995
Shrijal Acharya retweeted
Codex Computer Use is the most amazing technological capability I've ever personally witnessed.
When AI can do anything you can do on your computer...
24/7
No sleep
The question becomes:
> What's worth it to do on a computer?
29
14
366
32,217
What a thing this is. Woow
Introducing Claude Opus 4.7, our most capable Opus model yet.
It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back.
You can hand off your hardest work with less supervision.
ALT Claude Opus 4.7 Benchmarks
49
