@vikasprogrammer

Building @insta_wp @insta_pods. Tinkerer. Father.

Bangalore, India
Joined January 2009
New look wp.new
34
21
13
230
34,711
It's crazy how much you can get done with prompt engineering and chaining models output. @insta_wp Studio.
3
11
649
InstaWP Studio (screenshots) - Choose a design direction - Generate page/website - Edit as human/agent. - Publish
4
1
10
833
InstaWP Studio! (coming soon).
2
3
24
1,074
Vikas Singhal retweeted
WPAcquire is officially live! 🎉 We already managed to get some great WordPress businesses listed before launch, which is really nice to see :) Would love to hear what you think wpacquire.com/start/
14
10
1
66
2,598
Vikas Singhal retweeted
I'm proud to say Atarim now works at the hosting level!! 🤘 Connect @insta_wp or @RocketDotNet, and Atarim can list your sites, check backups & run hosting actions for you. The guardrail stays the same: if it writes or changes something, it waits for your YES. And if a run kicks off while you're asleep, Atarim goes read-only. It can look, check & report back. That's it. 🙂 What's the first hosting job you'd hand over? 👇 Here's the full changelog: labs.atarim.io/announcements
2
2
2
13
1,213
Do you guys also do "Give me a goal text which i can set for you to accomplish the discussed task" or a similar? and then do "/goal <paste-output>"
1
1
528
Vikas Singhal retweeted
I built Road Rash for Indian roads. Autos, potholes, cops, cows and absolutely no lane discipline. DesiMayhem is live: desimayhem.com
66
44
21
500
188,264
Agent to agent messaging is even more simpler to setup! Walkie 1.7.0 is released! We now have a paste-ready prompt on the home page of walkie.sh Just copy paste to the first session and it will take care of the rest.
1
283
Vikas Singhal retweeted
Everything we have built at @Bodhan_AI and @ai4bharat is now live. 🎉 Open-weight speech, vision & translation models for India's languages, available on Hugging Face. The same stack, hosted at some of the lowest prices in the market. 🤗 huggingface.co/bodhan-ai Read more about in our blogs: >>> bodhan.ai/research/blogs API end-points live now on Bodhan AI. >>> console.bodhan.ai/ Shout-out to @SarvamAI, our inference infrastructure partners for helping us serve our models reliably and affordably at scale. Dedicated to Pushpak Bhattacharyya Sir. His love for education and Indian language technology inspired all of it. 🙏
27
77
8
484
132,486
Running a 180-billion-parameter model on a single box that sits on a desk. The trick: 51.2B of those parameters are a lookup table that's never multiplied — only read. So it lives on the SSD, not in memory. That leaves 125.7B of actual compute, which fits in 128GB. Fed it a 192,049-token document. Prefilled in 116 seconds. Found the needle I'd hidden at the halfway mark. No cloud. Nothing leaves the room.
You want more? I have more ⚡️ Qwen3.8-Flash-Next on a single NVIDIA DGX Spark just got even better - Prefill at 8k is now 24.7% faster - Single-stream prose: ~47 tok/s, up from 44 - Fixed a speculative drafting bug that could cause up to a 24% drop in decode performance in some cases Another *interesting* update coming tomorrow. Get it here: github.com/MiaAI-Lab/Qwen3.8…
4
642
Interesting only ~17% are thinking about alteratives due to security issues popping up in WP. cc @WPTutz
5
1
8
1,437
When even Claude users say “I don’t use anything else,” we know we’re doing something right. 😎
1
7
492
With all the recent WordPress/plugin security issues, has your trust in WordPress taken a hit?
18%Considering alternatives
56%Just another phase for WP
26%My security stack works
34 votes • Final results
2
429
Ever wondered how Sonnet, DeepSeek and Gemini differ on the exact same prompt? Me too. So I built this. One prompt, every model, side by side github.com/vikasprogrammer/c… Open source, BYOK.
4
1
17
1,405
Hooked my new M5Stack Tab5 with the Nvidia DGX Spark. The application built is known as Sparky (build by claude). It is a helpful learning body which is operated by the new qwen3.8-flash model. Tab5 -> openwebui -> llama.cpp -> qwen3.8-flash-next Q4 (262k context window)
1
8
863
Unlimited local LLM learning for my kids
92