Vikram Bhalla

I’ve started a Substack! 🤓 Every week, I’ll be talking about all the ways I use AI to remove/reduce big and small frictions in my life and the lives of those around me (primarily @Kaju_Katri, who is my most patient guinea pig). It’s going to be about giving AI the stuff we hate to do, so we have more time and mental bandwidth to do the stuff we actually love. I’m going to try and keep it light, not too technical, and maybe even… fun?? gofrictionless.substack.com

Gautam Khattar row: People of Goa continue to defy polarisation tactics. But it no longer feels easy—Karanjeet Kaur @Kaju_Katri discusses in this week's column for ThePrint. youtu.be/UD4sLqWrAtA

Follow my Substack for weekly posts about how I'm using AI to do the things I hate, so I have more time and space to do the things I love: gofrictionless.substack.com

For years, every new app or platform I onboarded began the same way. I had to introduce myself, add my work context, connect the dots, and hope that somewhere along that convoluted process, my life actually got easier. Now, I no longer add my context to the tools. I can finally build the tools on top of my context. If that very specific but very tiresome friction resonates, and you're curious to know how I solved it, the details are in my latest Substack article, linked below: gofrictionless.substack.com/p/im-tired-of-…

1/5: At a Parshuram Jayanti celebration on 18 April in Goa, Gautam Khattar, who describes himself as “half-YouTuber, half-journalist”, went up on stage and bit off more than he could chew. He declared that women should not be granted 33 per cent reservation in Parliament.

5/5: For more, read #ThePrintOpinion by Karanjeet Kaur @Kaju_Katri 'People of Goa continue to defy polarisation tactics. But it no longer feels easy' theprint.in/opinion/goa-de…

@karrisaarinen I wrote about exactly this recently. I think there are plenty of things we could be using AI for — but thinking shouldn’t be one of them. gofrictionless.substack.com/p/a-creatives-…

A common dynamic I observe with AI: it feels most impressive when you don’t know much about the subject, don’t care or don’t have a clear idea of what the you want. This applies across design, code, legal, and more. If I don’t know code very well, every piece of code it writes feels very impressive. Once you know what something should feel or look like, it becomes almost impossible to guide AI there. And you definitely can’t one-shot it.

BJP wants to be the party of Nari Shakti. Here’s a list to make that happen - ThePrint columnist Karanjeet Kaur @Kaju_Katri explains youtu.be/XwRveIBoe-A

Read the full story here: gofrictionless.substack.com/p/how-i-accide…

I see your scope creep and raise you a project that started as a tiny chatbot — built to switch between AI models in the same conversation — and somehow became a personal AI operating system with 16 mini-apps, a three-tiered memory architecture, and a Telegram bot that texts my iPhone when I’m away from my computer. This might be my most ambitious side project to date. There’s too much to cover in one post, but for now, I just wanted to introduce you to Mimir. When you realise you can build anything, you end up wanting to build everything. The full, often embarrassing origin story is up on my Substack. Link in thread. 🧵

'BJP wants to be the party of Nari Shakti. Here’s a list to make that happen' Karanjeet Kaur @Kaju_Katri, columnist, writes #ThePrintOpinion theprint.in/opinion/bjp-wa…

Should you be nice to Claude? Here's what @AmandaAskell, philosopher and ethicist at @AnthropicAI, says:

@googlegemma For those asking, Gemma 4 runs surprisingly quick on my 3 year old iPhone 15. It's a very capable offline model if all you need is to chat. I wrote about it here, for anyone looking for a deep dive: gofrictionless.substack.com/ai-that-doesnt…

Check out this video on how to run Gemma 4 locally on an iPhone! It runs completely offline and handles long context, meaning no data plan, no API calls, and no monthly fees required.

‘AI tools can be governed eventually. But who will govern the men? Meta wants everyone to create AI images. Indian men chose misogyny’—watch this week’s column for ThePrint by Karanjeet Kaur @Kaju_Katri youtu.be/H7Xu-vdmMwY

‘AI tools can be governed eventually. But who will govern the men? Meta wants everyone to create AI images. Indian men chose misogyny’—watch this week’s column for ThePrint by Karanjeet Kaur @Kaju_Katri youtu.be/H7Xu-vdmMwY

It’s kind of surreal, sometimes, stepping outside the AI bubble and seeing just how ill-equipped most of the “real world” is. On the one hand, last week saw the reveal of Anthropic’s most powerful Mythos model, capable of identifying vulnerabilities in the world's most advanced and fundamental cyber systems. A model so powerful that Anthropic isn’t releasing it to the public, and instead working with governments and critical infrastructure organisations to help them find and fix weaknesses in their systems, some of which have lain dormant and undetected for years. And then, cut to real life… I had to pay my taxes, and this is the state of the official Indian tax portal. 😐 “We are not ready for what’s coming” doesn’t even begin to cover it.

Even though my official title is "Creative Director", most of my daily work tasks have absolutely nothing "creative" about them. But they always felt like the price I had to pay to get to do the thing I actually loved. If that resonates, I have some great news. You can still use AI in your work without using AI in your work. Sure, AI will keep getting better at more and more things. But we still get to decide what we let it take off our plate, and what we leave on. Because figuring that out forces me to think about the work I love, and where the line is for me between outsourcing labour and outsourcing judgment. After a few years of trial and error, I've landed on 6 simple questions that help me draw that line for any given task. Read all about it on my latest Substack post: gofrictionless.substack.com/a-creatives-gu… Is there a question I might have missed that works for you? I know creatives have strong feelings about AI, and I'm curious to hear other perspectives.

'Meta wants everyone to create AI images. Indian men chose misogyny' Karanjeet Kaur @Kaju_Katri, columnist, writes #ThePrintOpinion theprint.in/opinion/meta-w…

I agree completely, it's hardly a workhorse model for power users. But for a lot of "normies" whose primary use-case is still chatting and getting answers, Gemma 4 does prove quite worthwhile when you consider its advantages. I wrote about exactly this here, for anyone interested! 😬 gofrictionless.substack.com/p/ai-that-does…

I'm running Gemma 4 on my computer with Ollama. Unusable with Claude Code. It can't even load and execute skills, so I had to stop. But the model is pretty decent as a chatbot using the Ollama UI. I've been cross-posting questions across Claude and Gemma 4, and I can use Gemma's answers without any problems. I wish we had a better UI harness for the model (with projects, memory, etc.)

If you tell your $20-per-month AI BFF every personal, private, confidential thing with complete abandon, please know that those chats are sitting on someone else’s cloud server somewhere. There is another, safer, option for folks who use AI primarily to chat, draft emails, summarise documents, or write school reports. It’s called Gemma, and Google just released it. It’s everything you love about ChatGPT, without the privacy concerns or the monthly subscription fee. Read all about it in my first Substack post! You know where the link is. gofrictionless.substack.com/p/ai-that-does…