
CS Leader | Behavioral economics + systems thinking | Learning ML | Writing about retention, experiments | Building https://xitter.cf/t.co/eLKqfxWnE2, https://xitter.cf/t.co/hzVaGyxQSP, https://xitter.cf/t.co/TAWrC6APIs
Chennai, India
Joined January 2010
- Tweets31.2K
- Following751
- Followers3.3K
- Likes15.4K
Day 13 of switching to Android after 7+ years of exclusive iOS. Last year's attempt failed due to Apple watch+ work device policy + bulky phone (S25 Ultra).
So trying with Fitbit Air & Galaxy S26 (lighter flagship rn - still miss the GOATed Mi5).
Going good so far ♥️
Same. 6 years of using lounges, regular work & personal trips, 12+ trips in the last 12 months alone including 4 international, and I don't think we have used half as much, even combined a family. 106 within few months is the worst kind of abuse.
106! That's probably more than the lifetime number for my entire family combined!
BOB got this wrong. They should have closed the card and ended the entire banking relationship, just like HSBC does in such cases. Repeat offenders deserve to be blacklisted.
People often complain that we promote cards and they eventually get devalued.
But it's people who exploit benefits for personal gain, often by reselling, who contribute to these devaluations. Unfortunately, those genuinely educating others end up taking the heat too.
The irony? The people misusing the system remain invisible, while those openly sharing information become the face of the problem.
Lakshmi Narayanan G retweeted
Try Underdog: free, capable, private AI running locally on your device
New build and Saluki 27B are out. Under 8GB, beats Qwen 3.8 27B at tool calling
Like, repost, and comment one task you wouldn’t give a cloud agent to get an invite🦴
Today we're releasing Underdog Saluki 27B
Based on Qwen 3.8 27B, Saluki is almost 7x smaller than its full-precision counterpart while it keeps 96% of its benchmark performance and beats Qwen at tool calling
It's the best model <8GB that runs in Underdog
Saluki is available today under Apache 2.0. Try today in Underdog underdog.ai/saluki
1. ChatGPT $200 -> $20 (will return to $100 after using the $2500 worth credits)
2. Claude $20 -> $100
3. Considering Gemini AI Pro for Colab
4. Lovable $21 (expires Apr 2027)
5. Cursor Pro+ $60 (expires Jul 2027)
6. X Premium+ $33
7. Amp Megawatt $20 (might pause for few mths)
My current AI subscriptions right now 💸
1. ChatGPT - $100
Expires Oct 15. Not renewing it.
2. Claude - $20
Planning to upgrade to the $100 plan 🔥
3. SuperGrok - $30 for 3 months - 2 months left.
4. Grok Bot - Basically using Claude Opus 5.5, so this feels like a pretty good deal 😂
5. Gemini Pro - Waiting for Argon 👀 Currently not using it
My AI stack keeps changing every month
What AI plans are you paying for right now?
In many ways, Orkut was my first social network. Was sad when it shut down (even wrote about it: gln.me/my-orkut-life/)
Glad to see @orkut cooking up something new again. Excited for it!
orkut.com/en/?r=ORKT-VXH3-W3…
Classy
I read through it. I’m not saying you were wrong in your analysis. You were mostly right. And trust me we are aware of all of it. But there are just times when from our end we can do everything right. But outside factors or other circumstances can turn those upside down. It’s easy to analyse an end product or result. But it takes 2-300 people to make a film. It’s never easy. Never simple. And never something you can apply formulae or academics to. No production house or actor in world cinema intends to make mistakes or a bad film. In life has everything gone the way you intended ? So imagine an entity that operates with an internal team and multiple external teams. I’ll always stand by my team and if you want to blame someone you can blame me.
I,for one, love that they're moving Cowork to Cloud. Tasks getting interrupted when laptop lids are closed has kept me tied to my desk quite often at work for long running tasks. This solves for that
Replying to @GergelyOrosz
Hi! I work on Cowork. Probably no surprise, but we want to make our users as powerful as possible with Claude - and you're right that explaining and keeping up with the capabilities is hard. Affected users got both an email and an in-app notice, but I'm grateful for the feedback - I think this change was more surprising to people than I'd like it to be. This one will be a longer explainer, bear with me!
The "old" version of Cowork runs model inference in the cloud, executing tool calls in an Anthropic-provided VM we shipped to your computer. We added the VM for capability, safety, and security reasons - mapping in just the data you explicitly added to your session. People loved what they were able to do with Claude but didn't love the disk, battery, and performance cost of running the VM locally. Also, people didn't love that closing your laptop means the work stops.
The "new" version of Cowork runs model inference and the VM in the cloud. Each session gets its own sandbox, not sharing state with other sessions. When the VM needs something on the users' device (like a file), the desktop app is responsible for that file access tool call. To be clear, Claude will only access files in folders that the user explicitly added to the session, the same way Claude only saw files explicitly added to the VM in the "old" version. Sandboxes are destroyed when the session ends!
We think this solves a lot of problems we've heard about (like using Cowork from a phone, keeping work running, or getting all the same power without losing battery to the VM) - but we'll keep iterating on new feedback we hear!
Same feeling. Pre-ordered. Those are 3 microbrand items (all 3 are from different categories) that I have pre-ordered now. Even made full payment on one of them. Hope I don't completely forget and forgo my payment like I did in the past. 😂
Yaay! First transaction done via Apple Pay Axis card to reload my @akshayakalpa wallet. Love how fast @Razorpay started supporting Apple Pay🔥
If Sachin paaji is Team GPT, then I am too.
I’m excited to announce my partnership with OpenAI. We have some interesting things coming up, and I can’t wait to share them with you.
I’ve always been curious... One question usually leads to another, and ChatGPT has certainly encouraged that habit. But it has also proved useful when I’ve needed to get things done, like planning my last trip with the family.
That’s what I’m looking forward to with this partnership. Discovering what else I can do with ChatGPT, sharing what I find useful, and encouraging more people to give it a try with something they care about.
A question, an idea, something you’ve always wanted to explore. There are so many places to start.
Looking forward to exploring together.
#Partnership
Bankers job is infinitely more stressful, critical, and vital than ours. Such inane comments undermining that is absurd.
I don't want a burnt out banker handling our monies. But a 5 day work week isn't the answer. Banks have to figure out a way to have rotational shifts like IT does, to handle a 6 or 7 day week. They have plenty of profit to not do so.
Number of bank holidays apart from Sunday are nearly 15-25 days based on RBI list.
This is apart from the 36 days of SL-PL-CL that is mandated as per labour laws.
Corporate India employees have max 8 days of public holidays + 2 RH & regular leave credits as per labour laws.
Some MNC’s give only 30 days leave credits per year including all public & national holidays.
Banking is a public service & cannot be mandated with 5 day working.
They have already 2nd & 4th Saturdays off, if they want only 5 day work weeks then their bank holiday lists should be curtailed to max 8 days in a year
Had a wonderful stay @TajHotels Fisherman's Cove last weekend for my birthday at a ~71% "discount" (lacking a better word) - paid 35,430 for a list rate of ~1.22 lacs (2 nights).
EPM Epicure FNA, member discount, Amex Taj voucher, on-spot upgrade discount & EPM iShop vouchers.
Been exclusively using Opus 5.5 on medium since last night on the $20 plan.
Only about ~52% of weekly limit gone. Not sure about token value since I've been running on the cloud VM they give - it has shipped a ton of work Astra Light struggled with.
Impressive.
Tried @claudeai for a side project after ~6 months. Pleasantly surprised with the improvements.
Here's what I loved:
- How far Opus 5.5 goes on the $20 plan
- Design sense (something GPT, including Astra, is still struggling with)
- Auto-resume on the 5-hour limit reset
I'm 10k away from hitting 7.5L/Gold tier on Atlas. Card's main use was ITC for me. Renewal next month.
Still worth keeping & hitting it or should I just give it up & double down on Infinia + EPM? @jassneetsingha @suritalreja @imYadav31 wdyt?
Tried @claudeai for a side project after ~6 months. Pleasantly surprised with the improvements.
Here's what I loved:
- How far Opus 5.5 goes on the $20 plan
- Design sense (something GPT, including Astra, is still struggling with)
- Auto-resume on the 5-hour limit reset
Tried @voicenotesai's dictation feature today. Mind blown. Better than @WisprFlow or @usemonologue for my accent. Great to use too.














