Angehefteter Tweet
vikas sabbi
4K posts
vikas sabbi
@vikassabbi
In pursuit of greatness • prev @spawn @betteryou_app @ibm
Minnesota, USA Beigetreten Haziran 2011
741 Folgt1.1K Follower
today i turn 29 🎉
today is also 4/20
so i'm dropping a new song.
it's called... White Mirror
enjoy!
English
there has never been a good time to be a cybersecurity engineer.
with vibe-coded apps flooding the internet + AI-powered hackers.
the hottest job in tech soon will be Chief Security Officer.
Guillermo Rauch@rauchg
Here's my update to the broader community about the ongoing incident investigation. I want to give you the rundown of the situation directly. A Vercel employee got compromised via the breach of an AI platform customer called Context.ai that he was using. The details are being fully investigated. Through a series of maneuvers that escalated from our colleague’s compromised Vercel Google Workspace account, the attacker got further access to Vercel environments. Vercel stores all customer environment variables fully encrypted at rest. We have numerous defense-in-depth mechanisms to protect core systems and customer data. We do have a capability however to designate environment variables as “non-sensitive”. Unfortunately, the attacker got further access through their enumeration. We believe the attacking group to be highly sophisticated and, I strongly suspect, significantly accelerated by AI. They moved with surprising velocity and in-depth understanding of Vercel. At the moment, we believe the number of customers with security impact to be quite limited. We’ve reached out with utmost priority to the ones we have concerns about. All of our focus right now is on investigation, communication to customers, enhancement of security measures, and sanitization of our environments. We’ve deployed extensive protection measures and monitoring. We’ve analyzed our supply chain, ensuring Next.js, Turbopack, and our many open source projects remain safe for our community. The recommendation for all Vercel customers is to follow the Security Bulletin closely (vercel.com/kb/bulletin/ve…). My advice to everyone is to follow the best practices of security response: secret rotation, monitoring access to your Vercel environments and linked services, and ensuring the proper use of the sensitive env variables feature. In response to this, and to aid in the improvement of all of our customers’ security postures, we’ve already rolled out new capabilities in the dashboard, including an overview page of environment variables, and a better user interface for sensitive env var creation and management. As always, I’m totally open to your feedback. We’re working with elite cybersecurity firms, industry peers, and law enforcement. We’ve reached out to Context to assist in understanding the full scale of the incident, in an effort to protect other organizations and the broader internet. I also want to thank the Google Mandiant team for their active engagement and assistance. It’s my mission to turn this attack into the most formidable security response imaginable. It’s always been a top priority for me. Vercel employs some of the most dedicated security researchers and security-minded engineers in the world. I commit to keeping you updated and rolling out extensive improvements and defenses so you, our customers and community, can have the peace of mind that Vercel always has your back.
English
Claude Design is one of the best 🔥 It created a full design system which is insanely good and you can import that directly to ClaudeCode.
Only thing that sucks is the limtations 😕 @bcherny can you please increase the rate limits for design too?
Claude@claudeai
Introducing Claude Design by Anthropic Labs: make prototypes, slides, and one-pagers by talking to Claude. Powered by Claude Opus 4.7, our most capable vision model. Available in research preview on the Pro, Max, Team, and Enterprise plans, rolling out throughout the day.
English
Am I the only one having a good experience with Opus 4.7?
I still vastly prefer Codex for most things but Opus is absolutely nailing every UI task I give it.
English
@fdotinc canopy update!
talked to dino fans across the us, india, philippines. showed the product, got feedback, pushed updates.
we got merch companies that are reaching out will finalise that soon or later.
vikas sabbi@vikassabbi
update we are 257k now, we will reach 400k by end of canopy.
English
@akothari you should listen to her, she’s amazingly talented
youtu.be/tdVqEcdxSDM?si…
YouTube
English
Recently got into Hindustani classical music and I’m really enjoying this playlist. It’s a great introduction to both the legends and some contemporary artists.
open.spotify.com/playlist/37i9d…
English
Left or right?
For the article...
Need your help here. I like both...
English
vikas sabbi retweetet
vikas sabbi retweetet
WOZCODE just dropped a free Claude Code plugin and it's making sessions 30–40% faster while cutting costs by up to 55%.
Two commands to install. Works with your existing subscription.
Here's why this matters 👇
Claude Code is powerful but it burns through tokens like crazy.
WOZCODE sits on top of your existing Claude Code setup as a plugin. No new subscription, no config changes, no data leaving your machine.
What it actually does:
→ 25–55% cheaper sessions (less token waste, fewer unnecessary round-trips)
→ 30–40% faster on most tasks, up to 10x faster on database operations
→ 78% on TerminalBench 2.0 vs 58% for Claude Code alone
That last one is wild. A plugin scoring 20 points higher than the base tool on real-world coding benchmarks.
They ran head-to-head tests across 3 categories:
1. App Build (10 sequential prompts. Pages, styling, interactivity, responsive design)
2. Database Queries (10 queries against a live 11.5M+ row clinical trials database with 68 tables)
3. Code Edits (15 prompts. Bug fixes, refactoring, type safety, testing on a real Express + TypeScript + PostgreSQL repo)
WOZCODE completed the same work with fewer round-trips and lower token consumption across all three.
The install is dead simple:
Two commands in your Claude Code CLI. That's it. No proxying, no data interception. Everything runs locally within your existing Claude Code session and Anthropic's infrastructure.
Works with Claude Code CLI, VS Code, and Cursor.
If you're hitting your token limits before finishing real work, this is worth trying.
English
vikas sabbi retweetet
English
@NoahHirshon @vikassabbi @WithWoz1 @ycombinator +1 If you can really improve Claude, go work at Anthropic. This is trash.
English
@NoahHirshon @WithWoz1 @ycombinator not necessarily, there are plenty of strong benchmark results stay niche unless they get picked up by bigger accounts media or the open source crowd at the right moment. github traction usually reflects distribution and timing, not just whether a claim is true.
English
@vikassabbi @WithWoz1 @ycombinator if that was true wouldn't this be easily the number one github repo of the day or be known outside of this post?
English
@NoahHirshon @WithWoz1 @ycombinator those are the real benchmarks you cannot fake that, if you still don’t believe it you should give it a shot
English
@vikassabbi @WithWoz1 @ycombinator i guess i just don't believe that one plugin improves benchmarks by 20% on any model with any plugin lol
English
vikas sabbi retweetet
😱 A CLAUDE CODE PLUGIN JUST DROPPED. 40% LESS TOKEN WASTE. SESSIONS 30% FASTER.
It is called WOZCODE And it FIXES THE ONE THING that kills every Claude Code session. You can INSTALL IT FOR FREE.
Here is what it actually does.
An actual plugin that sits inside Claude Code and silently optimizes every single session automatically.
The problem nobody talks about:
Every turn you make in Claude Code, the ENTIRE conversation history gets re-sent to the model. All of it. Every single time.
By turn 100, each message costs 7x more tokens than turn 1. By turn 300, it is 20x. You are not running out of quota. You are burning through it on context you already have.
WozCode fixes this at the infrastructure level:
Token Compression:
→ Auto-compacts context at 50% window capacity instead of waiting for overflow
→ Never lets dead context carry forward into new turns
→ Strips raw tool output noise from MCP calls before it hits your context window
→ Community-reported result: 68% fewer tokens burned per session
Session Intelligence:
→ Monitors token consumption per turn in real time
→ Automatically rotates sessions before they become expensive
→ Injects compressed context into the fresh session so you never lose your place
→ Blocks sessions from continuing when burn rate crosses your set threshold
Speed Engine:
→ Reduces time-to-first-response because smaller context = faster processing
→ Sessions run 30 to 40% faster on average
→ Works across Claude Code CLI, VS Code, and JetBrains
The numbers people are reporting:
→ 64,900 tokens saved over two days of normal use
→ Sessions that used to hit limits in 2 hours now last 4 to 5 hours
→ Pro plan users going from 10 to 40 Claude Code prompts every 5 hours to 55 to 70
→ Works with every Claude model including Opus, Sonnet, and Haiku
The core innovation:
Most people try to fix token waste with prompts. "Be concise." "Summarize this." That does not work. The problem is not your prompts. The problem is that Claude Code re-reads everything, every single time.
WozCode intercepts at the session level. It manages what goes into context, what gets compressed, and when to rotate. You just keep working. It handles the rest.
Install in 30 seconds:
→ Works with Claude Code CLI + VS Code + JetBrains
→ Run one command to install
→ Runs locally, no data leaves your machine
→ Zero configuration needed to start seeing results
The wildest part? The context window problem has been sitting in plain sight since Claude Code launched. By turn 300, you are paying 20x per message compared to turn 1. Nobody was building a fix for it at the plugin level.
Now someone did.
🔖Save this before your token budget runs out again
English
vikas sabbi retweetet
I think I've been overpaying for Claude Code this whole time.
Ran the same task twice today. Once normal. Once through this plugin called WOZCODE.
Stock Claude Code: 236 turns, $14.38. WOZCODE: 76 turns, $6.61. Same Opus 4.7. Same prompt.
English
English
@WithWoz1 @ycombinator @grok what's the value proposition of this plug-in instead of claude? Is there an extension for a vs code? Doesn't make sense. It'll use less tokens and be more efficient. And will I be more efficient because of this?
Warren, OH 🇺🇸 English
vikas sabbi retweetet
Someone built a plugin that cuts Claude Code's cost by 54% on the same task, same model.
Not a different model. Not a different harness. Just smarter tools underneath.
It's called WOZCODE and the benchmark numbers are genuinely CRAZYYYY.
English
Stop burning cash on Claude tokens.
Save that money for your summer vacation instead 🏖️
Woz@WithWoz1
Claude likes wasting your tokens. You don't have to let it. Install WOZCODE for free. A plugin for Claude Code that makes it: → 25–55% cheaper → 30–40% faster → +20% better on benchmarks Installs in 2 commands. Works in your existing workflow.
English
@WithWoz1 @ycombinator why is Y combinator reposting this shit lol
English