BhavdeepSethi

Same, I like the general idea of that post. But the structure and verbosity made it such that I wouldn't want to share it with others.

I'd really like to read the Post Mortem doc for this. I'm still shocked how they allowed this to happen in production.

I got an ad recently on chatgpt for asking a question regarding bleach. The recommended product in the answer wasn't the same as the ad though. Ad was also at the bottom. Wondering how long before they go the Google route, and show top 5 links with ads before answering anything.

I wonder if GH charges for the runners during their downtime. Last week lot of them would retry multiple times and then fail.

Most of the folks that are talking about this are the ones who work independently and work on greenfield projects (especially tooling related). The cost of making a mistake there is so low. I've used it similarly and it's absolutely amazing. Though I still use a mix of agents and code myself in my regular 9-5 job.

I've yet to see examples of folks using this in a team of 4+ folks working together in a production env with users, and just using AI for their regular development.

Claude code creator only using claude code doesn't count. That's more like dog-fooding.

Cursor is an IDE.

Pachinko has been one of my favorite reads for a long time. If you enjoyed that, you'll also like Green Bone Saga by Fonda Lee.

Merry Christmas to you and your family!

First, allow anyone to sign up for the blue check mark for $8. Verified accounts loses its value and Twitter gets flooded with fake accounts and foreign run accounts. Now, try to fix it with by showing users country of origin. Now, these users will try to figure out ways to bypass it.

From what I've learned, be as explicit as possible when users enter dates.

We had to change our date of birth field in the user sign-up to three separate "Month Name","Day" and "Year" drop downs, since so many people made mistakes (fat finger/ swap month and day) from the "MM/DD/YYYY" field, and would then send support ticket to update it.

Funny since they have a tender offer that hits their accounts on Oct 7.

I don't think they meant desktops in the literal sense. Laptop with/without monitors is effectively considered desktop now (compared to mobile web/apps).

A lot of folks took the bet on CloudKitchens because of Travis. Even after raising $850m, CloudKitchens has never ran liquidation event for employees afaik. I wonder how they attract/retain talent in that scenario.

My experience has been quite different. With Cursor/Claude code, I've ended up writing full fledge solutions (running cli/web servers with loggers and unit tests for each functionality). We're talking crawlers, cab booking service like uber, search engines with seed data. All within the hour.

Definitely not insane. Ironic is the correct term. The field is evolving, a lot of these companies talk about replacing outdated practices using AI. Asking software engineers to not use their own tools to solve problems falls under the same bucket.

It's insane how many of these new "AI" companies don't let you use AI or even your own IDE for coding interviews. And most questions from such companies are LC type problems so they know any AI tool can one shot it.

You're absolutely right! The degraded model quality finally pushed me to stop paying for the max plan. Still on the Pro for now.

It's funny that the MIT paper HN is trending higher than this post, so it propped up before I read this article.

I don't believe they did. This acquisition was by Flipkart, the poster child startup in India, who had a very high bar for hiring. They wanted to interview the non-founders to make sure they met the standard. The founders said you get all or you get none. To be fair, it was a small team of 6-8 employees, so I doubt Flipkart cared. :)

I went through an acquisition very early in my career, and for the longest time I believed it was the best outcome for everyone. Over time, I realized that my naive belief was purely due to the founders going way above and beyond to make sure each and every employee (including folks doing just data entry) got a good outcome (accelerated vesting, significant equity in new company, top of the band pay, etc.). It made me realize that if you ever want to work at a start up, bet on the founder, rather the company. Even with mediocre outcomes, you'll end up ahead in the long run compared to folks who're just looking out for themselves.

Yep, didn't intend to downplay your claims. Would love to see a video in action to learn more.

> “commit everything in logical chunks” and Claude handles the entire flow—staging changes, writing meaningful commit messages, pushing, opening PRs, watching CI, and fixing any CI failures. When builds break, it analyzes the errors and patches them automatically. It’s also extremely good at resolving merge conflicts.

Is this marketing hype for the post or does it actually do all this? I know it's relatively new, but I haven't been able to find a resources to do such a setup? Only posts glorifying that it's capable of such tasks. If anyone else has had better luck with such complex tasks, would love to find out more.

Surprised they didn't answer if they tried using AlphaEvolve to improve AlphaEvolve!

The universe is a cruel, uncaring void. The key to being happy isn’t a search for meaning, it’s to just keep yourself busy with unimportant nonsense, and eventually you’ll be dead. - BoJack Horseman

I thought this was satire before I looked it up. I mean, it makes sense but it's still so hard to believe.

This is insane to read. You don't get brownie points for what you're wearing, but deducting points for someone trying to impress folks by dressing more formally than what you're typically used to?

15 years ago, the first start up I worked for provided APIs for music streaming in India. One of the founders who managed all infra was in US, and so the servers (bare metal) were in LA. I still find it amusing, that it was cheaper (and faster) just to fly to India, buy bunch of portable hard drives, upload the media, fly back to US and upload the data to the file server, than uploading the media directly from India to the US server. Obviously only applies when data is in order of TBs. Later saw the same thing with AWS Snowball and Snowmobile.

Elon made this popular with Twitter/X. Offered a voluntary exit in the "Fork in the road" email with a deadline. The message was quite similar, that if you're not aligned with the company's [new] vision, you can choose to leave with severance.

It seems better than doing random layoffs, no?

HuggingFace demo: https://huggingface.co/spaces/Qwen/Qwen2.5-Max-Demo

Source: https://x.com/Alibaba_Qwen/status/1884263157574820053

I honestly feel anyone comparing AI to Web3 is absolutely delusional. Web3 hasn't had a single mass use-case except being a speculative asset. AI is already being used in lot of real life applications (radiology, law, self driving, etc.)

The negatives mentioned for AI here can be used for any technology application which reduces manual labor. AI is gonna enhance your job, or going to displace to a better job for you. Why do you wish to continue doing the work that technology can do 10x better?