033 • 12 August 2025

Contents of this Newsletter

🛠️

Cracking Mobile Automation

⏰ : 5 min

📓

Blog Picks

⏰ : 3 min

👨🏻‍💻

New Team Members

⏰ : 7 min

🗞️

Weekly AI News

⏰ : 9 min

MAIN STORY

When Software Feels Like Magic

Technology is fragile. Even the best tools can stumble when complexity shows through. The ones we truly love (the ones we trust) are those that disappear into the background and simply work. That’s the standard we chase at Quash. Last week, for the first time, we hit it.

It started quietly. Our CEO, Ayush had been off the radar for days, little did we know he was solving something that was long overdue; our execution engine. When he finally called us together, we expected a routine internal test. Instead, we saw something that stopped us in our tracks. The engine ran with a speed and precision we’d never seen before, moving like a real user, adapting mid-flow, making its own decisions. It was the kind of moment that makes you realize you’ve cracked something important.

The Breakthrough

This wasn’t automation replaying clicks: it was interpreting, weighing options, and choosing the most relevant actions without being told. It adapted to context shifts, prioritized intelligently, and executed without hesitation.

Ayush called it a “quantum leap in mobile automation.” This is automation that blends cognitive decision-making, dual-image visual analysis, workflow awareness, and contextual prioritization. It doesn’t just follow instructions but it understands them.

The room felt different afterward. This wasn’t a small step forward. It was a pivot point for the company, finally, a moment when the vision became tangible.

From Breakthrough to Blueprint

The demo made something clear: our execution pipeline is stronger, faster, and more reliable than the generation-first path we’d been pursuing. Having two separate pipelines was creating friction, so we’re folding them into one cohesive system. The goal being simpler architecture, fewer dependencies, and higher quality across the board.

We’re also exploring a bold shift: we’re experimenting with new ways to center releases in the process. The kind that could cut down on prep time while giving teams sharper signals about readiness. What if each release could define its testing flows on the fly, delivering a readiness signal unique to that build? Maybe it’s the future or it’s just an experiment. Either way, we’re pushing boundaries.

The Beginning of the End

The energy in the office has changed. Product and engineering are closing gaps, validating every edge case, and refining the architecture. Growth is in working beside these teams, crafting the messaging without spoiling the reveal. In the midst of this, the team itself has evolved. Some familiar faces have moved on, and new talent has joined the fold. It’s been a quiet shift, but one that has added fresh energy and new perspectives just when we need them most.

We’ve seen what’s possible. The breakthrough we witnessed has changed our vision for what Quash can be. The team is focused on polishing every last detail, determined to deliver an experience that truly redefines quality assurance. Everyone is locked in, just trying to complete and ship this as soon as we can. We sure are excited for you guys to see what's yet to come.

When launch day comes, it will be a system that thinks like a tester, moves like a real user, and makes QA feel effortless in the best way possible.

We’re not there yet, but the countdown has begun.

Check out our Recent Blogs

Welcome To The Team!

This week we’re happy to welcome 3 new members to our team: Khushi, Abinav and Shreyansh.

Khushi

Fun Fact: If it involves dogs, food, or a flight to somewhere new, I’m in. Add anime and nighttime energy? That’s my whole personality right there.

Hi, I’m Khushi, joining Quash as a Marketing Intern. I recently completed my Master’s from Aston University, Birmingham, with a passion for creative strategy and brand building.

Abinav

Fun Fact: Selectively chaotic, absolutely the best in every game I’ve grinded, forever hungry, and always down to drop in on anything spontaneous.

Hi, I’m Abinav, and I’ve joined Quash as a Founder’s Office intern. I thrive on stepping out of my comfort zone, picking up new skills fast, and my ADHD keeps me curious, energetic, and always moving.

Shreyansh

Fun Fact: I unwind by coding, exploring the world alone, and battling both F1 strategies and chess opponents.

Hi, I’m Shreyansh, a fullstack developer with experience in building scalable web applications and AI-driven solutions. Looking forward to breaking prod as a frontend intern at Quash.

News

This Week in AI

Gpt 5 Rolls Out, So Does Backlash
OpenAI announced GPT‑5 and GPT‑5 Pro, positioning them as significant leaps in reasoning, coding, math, and safety via a new “safe completions” paradigm. Within days of launch, Reddit and social channels filled with complaints that GPT‑5 felt “sanitized,” less warm, and disrupted workflows, especially after legacy model options (like GPT‑4o) were hidden or removed by default. OpenAI CEO Sam Altman admitted a model-switching bug degraded performance and promised fixes, including better routing, higher rate limits, and an optional “thinking mode.”

OpenAI released open-weight models: gpt‑oss‑120b and gpt‑oss‑20b
First open-weight LLMs from OpenAI since 2019, under Apache 2.0. They target advanced reasoning and tool use, offer adjustable reasoning effort, support full chain‑of‑thought for debugging, and are intended to complement hosted models. Downloads are available for local and on‑prem use. Independent analyses highlight that 120B approaches small OpenAI reasoning models on internal benchmarks, and 20B is usable on commodity laptops; both use MoE with relatively small active parameters per token.

NVIDIA Unveils Cosmos World Models & Infrastructure for Physical AI
NVIDIA has launched new Cosmos world models, which are AI models trained to understand and simulate physical environments aimed at robotics, industrial automation, and other real-world applications. They also introduced infrastructure tools to make deploying AI in physical systems faster and more reliable, pushing AI’s role beyond software into hardware-driven industries.

All Newsletter Issues | Security | Contact Us | Blogs | Test Case Generator

Unsubscribe · Preferences

We Cracked Mobile Automation 🛠️ + The Real Story on GPT-5