@championswimmer@_pratikpakhale Something tells me your tweet is going to set in motion a series of events which will lead to them changing this unsustainable pricing/model 🙂
@_pratikpakhale They give you 1500 sessions. Each session is like 25-30 LLM calls.
Small tasks can be done in one session. Larger tasks take 2-3 sessions.
They don't charge you by tokens and don't rate limit API calls.
Most people are sleeping on the fact that Copilot CLI is just better than Claude Code or Codex subscription.
1. You switch between codex and claude in same subscription
2. $39 sub is great value (3-4x usage than $20 plans)
3. the CLI is so good! no flicker and /slash commands work even when model is working. Unlike CC or Codex.
Finally listened to the Dario interview and it blows my mind how his entire worldview around AI seems to be predicated on an assumption nobody is going to YOLO upload a tarball of frontier model weights to the Internet.
I made an app to turn web analytics into a Severance episode.
Every page on your site is a department.
/pricing is a department.
/blog is a department.
Your visitors are innies now.
When they visit your site, they arrive through the elevator.
When they navigate to another page, they walk to the next department.
When they bounce, they walk out.
It's called MACRODATA: macrodata.datafa.st
I'm trying to figure out where to live next, and one big consideration is the climate. So naturally I made a tool that represents monthly average temperatures for cities as 3D rings so they can be compared more easily. Check it out: awjuliani.github.io/weather-explor…
Someone ordered eggs on Instamart and only one came cracked.
Instead of just reporting it, they opened Gemini Nano and literally typed:
“apply more cracks.”
In a few seconds, AI turned that tray into 20+ cracked eggs — flawless, realistic, impossible to distinguish.
Support took one look at the “proof”…
processed a full refund…
and moved on.
Just pause and think about that.
Our refund systems were built for a world where photos were trustworthy.
But now they’re up against 2025-level AI —
and they’re getting absolutely destroyed.
If even 1% of people start doing this,
quick-commerce unit economics won’t just suffer —
they’ll implode.
AI isn’t the villain here.
The real problem is verification systems stuck in the past.
Welcome to the era of AI vs AI.
@cursor_ai can you give us programmable buttons/one-click-actions that can be run on files?
Write a prompt and link it to a button, button shows up on UI based on context variables used in prompt.
Eg: Button with prompt "Write tests for $file" shows up on file views in editor.
@vinayw@yigitboyar I left Google to go back to school. Isn't that what everyone does?
(@chethaase/so-long-and-thanks-for-all-the-bytes-02a4ef972f65" target="_blank" rel="nofollow noopener">medium.com/@chethaase/so-…)
Yesterday was my last day at Google.
What started as a 2 year Green Card journey became 11 amazing years & meeting incredible people.
So many great memories from creating Jetpack to adopting Kotlin. 10/10!
It is time to move on though, try new challenges and learn new stuff!
@yigitboyar If I were still there, I'd be very sad to see you go. Now I'm only happy for you.
Good luck with whatever comes next! I've heard that there is life after Google. And homework. Lots and lots of homework.
Excellent.
My workflow is now this:
Start with every file committed. Split screen.
Left side is Windsurf to build new features.
Right side is Cursor to build test suite to 80% code coverage (on previous features). This activity happens while I wait for Windsurf to finish.
When new feature is done, windsurf is really bad at git commits, so I ask cursor to @commit diff tool, and have it group files for commits with great messages. It's very easy to see if there were any bugs with the git tab. The diff is right there. So I do spot checks.
Then in the background I have Gemini 2.5 Pro in aistudio and O3 w/ ChatGPT w/ their 1M context windows, to look at the project from 30K ft. Since they both have large context windows, I use ChatGPT to debug and Gemini to help plan new features. Often while WS and C are doing their thing.
Then I have Claude to write small focused scripts for little stuff here and there.
A game changer is I vibe coded a script that grabs all the md, yaml and toml files (since Gemini won't let me upload them) and the code signatures from py files (inspired by @simonw's symbex) to generate an llm.txt file every time a new feature is deployed.
I used to upload my whole codebase to gemini and ask it to find errors or stuff like that. But now the llm.txt is much smaller and does the trick. Sometimes I tell Gemini to let me know if it wants to see any py files.
Also, to debug or just generally make sure something was done well, within that context of the llm.txt, upload the changed .py files and ask if Windsurf/Cursor wrote them well.
It's an orchestra of AI employees, and quite involved, but dang if it hasn't tamed the AI coding problems that have plagues the last few months.
we’ve found forking chats to be useful when you want to go back and explore new paths while preserving old ones. here we’re forking a chat from the middle of a conversation and continuing with a different model. what do you think, how can we make this better?
We love @NotionHQ, and now you can share your Granola notes to Notion with just a single click 🏃♀️➡️
Go to Granola's settings to turn it on. Kudos to @nyalsadiq & @_tikhon for making this happen 💪
Someone installed a solar powered Android device with Shazam constantly open to track the songs that are being played on the streets of San Francisco.
He calls it "Culture Surveillance".
Boy do I love experiments like these.
Issue #6 of Dispatch goes out tomorrow morning. I have no doubt that it's the best one yet. Here's a quick preview of what to expect.
Subscribers also get good karma for the next 5 years. Easiest "W" of your life 😂
it'd be NICE if someone does a proof check like this in hyderabad too
thankyou very much in advance
i miss delhi ke chole bhature SO MUCHH🥺
(note: I'll believe only those who've had CB in delhi before)