Ravi Mohan

Ravi Mohan

@ravimohan.bsky.social

Generalist Software developer based in Sydney, working with c# | angular | aws at work and c# | nextjs | fly.io for the side. Technology Agnostic and love to play with all languages and tools. https://github.com/sravimohan

420 followers499 following363 posts

View full thread

Even the most explicit FAQ, they are trying to weasel this. Sure, storage is getting expensive but be upfront about it and don't treat customers like idiots.

Also, this is miscalculated and will hurt their copilot push. Lot of the MS 365 customers are there for the storage. Office apps are just a bonus.

Here is how they decided to word the change "More flexible cloud storage: Your cloud storage allowance will change from 1 TB per person to 2 TB of shared storage so everyone on the plan can use the free space while keeping their own accounts and files private." We are doing you a favour guys.

AWS Sydney region has run out of capacity for RDS db.t4g.medium and is affecting existing instances. This is wild, I thought the cloud had unlimited capacity 😐

I called this couple of months ago. code.claude.com/docs/en/clau... are officially available, they let you keep Claude Code working, even when your laptop is closed. bsky.app/profile/ravi...

code.claude.comUse Claude Code in the cloud - Claude Code DocsRun Claude Code sessions in the cloud from your browser, phone, desktop app, or terminal, move them with --cloud and --teleport, and auto-fix pull requests.
Ravi Mohan@ravimohan.bsky.social

Hot take 🔥: agentic loops don’t care about your commute. My laptop closes its lid 💻😴, my wifi wobbles on the train 📶💀, and suddenly my AI agent is just… vibing in a half-finished thought. Laptops are becoming glorified remote controls.

AI is allowing people to be cross role boundaries. But I am seeing some tension, with team members wanting to hold on to their domain and being protective. I predict lot more generalists and less specialists in the future.

Hot take 🔥 for today. With agentic development, it makes sense to assign developers epics rather than stories. They can just get the whole thing done end to end without the friction of per story red tape.

It is much easier for two developers to work on two independent epics then to share an epic and try to prioritise and organise stories. However, knowledge transfer would be the issue.

Claude having a moment right when my weekly limit reset's is not helping. I’m basically living inside that XKCD outage graph — everything fine until the exact second I need it.

Hot take 🔥: agentic loops don’t care about your commute. My laptop closes its lid 💻😴, my wifi wobbles on the train 📶💀, and suddenly my AI agent is just… vibing in a half-finished thought. Laptops are becoming glorified remote controls.

The old dev loop tolerated interruptions — you’d context-switch back in seconds ⏱️. Agentic loops don’t forgive that. A dropped connection mid-run isn’t a pause, it’s a lobotomy 🧠❌ for your agent’s train of thought. 2/3

So we’re heading toward always-on cloud dev boxes ☁️🖥️: 24/7 reliability, agent chugging away while you sleep/commute/touch grass 🌱. Your laptop’s new job? SSH client. That’s it. That’s the tweet. 🚀 3/3

Got an email titled “GitHub Code Quality GA is July 20.” I almost deleted it — thought it was just another GA announcement. Then I actually read it and went: hold up… they’re auto‑starting billing? $10 per active committer/month + AI usage. What!?

View full thread

And the doc doesn’t show any way to see which repos even have Code Quality enabled. It literally reads like: “open a repo → disable → repeat.” It feels like GitHub expects admins to go on an expensive Easter egg hunt.

I’m not against paid features. I’m against starting auto‑billing for a feature which was free. This needs to be disabled by default until the admin re-enables it with clear understanding of the cost of the service.

Got an email titled “GitHub Code Quality GA is July 20.” I almost deleted it — thought it was just another GA announcement. Then I actually read it and went: hold up… they’re auto‑starting billing? $10 per active committer/month + AI usage. What!?

The email links to a doc on how to disable Code Quality… but it only shows how to turn it off per repo. No mention of an org‑level option at all. So, at first glance it really looks like you have to click through every repo manually. But you can turn off this feature at org level settings.

AI coding tools have changed how I work. I barely open IDEs now and I did not renew my #Rider license. That money is better used for my AI budget. It makes me wonder what this shift means for the future of IDEs.

If you are using Claude, chrome-devtools-mcp. The MCP tool results stay in context for the rest of the session. /compact to flush them. Or do the devtools troubleshooting in a throw away session.

Ravi Mohan reposted

Well this is massive: - Every non-US company sees how they can be cut off from US vendors with a snap of a finger. Non-US vendors getting free advertisement - What does this mean for US AI labs’ offices and employees outside the US? Reads pretty draconian What a mess

Ravi Mohan reposted

Replying to @gergely.pragmaticengineer.com

And yes, the potential impact could big even for US companies who employ even a single non-citizen. Anthropic shut down the model for everyone for now, but what the order asks for seems like a nightmare to implement without drastic Chinese walls inside a co... bsky.app/profile/kaip...

dogfood@kaipahl.bsky.social

Re: Every non-US company sees how they can be cut off from US vendors Are you not underselling the impact? Anthropic is writing: „any foreign national, whether inside or outside the United States, including foreign national Anthropic employees"

View full thread

The promise is "vendor-agnostic — apps built with Vite are meant to run anywhere." That's a direct message to the ecosystem: we're not making this a Cloudflare-only story. Whether that holds long term is the real question. 7/8

View full thread

Vercel has deep roots here too — they fund Turborepo, SWC, and have historically sponsored Vite. Their whole business is built on deploying Vite-powered apps. Cloudflare just put its flag in the same ground. 5/8

Vercel bets on the framework layer (Next.js). Cloudflare is now betting on the tooling layer underneath. Two different strategies for owning the modern frontend stack. Neither can ignore the other. ⚔️ 6/8

View full thread

The reassurances: Vite stays MIT. Stays vendor-agnostic. The same Vite team continues stewarding it. VoidZero members are all joining Cloudflare and continuing their work. Nothing changes today. 3/8

But here's what's interesting — Cloudflare is also dropping a $1M open source fund for the Vite ecosystem. Plugins, tooling, core team stipends, faster security audits. That's real money with real intent. 4/8

VoidZero was the company behind Vite, Vitest, Rolldown, and Oxc. That's not just a build tool — that's the foundational layer under React, Vue, Svelte, SolidJS, and most modern JS frameworks. Cloudflare now has a seat at that table. 2/8