A 22-year-old intern fixed a bug that 3 senior engineers couldn't solve for 6 months.
Saved the company $2M. They gave him a $500 gift card and didn't convert him to full-time.
He posted the solution on GitHub. A startup saw it. Hired him at $140k.
The 3 senior engineers who couldn't fix it? They're still there. Still getting paid. Still breaking things.
Talent gets punished in corporate. Mediocrity gets protected. That's not a bug in the system that's the feature.
Unpopular opinion on agentic workflows:
90% of the time, you don't need an autonomous agent. You need a standard, deterministic script that runs your tools sequentially, collects the output, and hands it to an LLM for final synthesis.
I must admit that nothing about computers, since I'm in love with the field, was so uninteresting as the Javascript different fashions, waves, frameworks, rewrites, hypes. And I'm one that loves almost every shit programming related.
Mistakes happen. As a team, the important thing is to recognize it’s never an individuals’s fault — it’s the process, the culture, or the infra.
In this case, there was a manual deploy step that should have been better automated. Our team has made a few improvements to the automation for next time, a couple more on the way.
UNIX was built in one month.
Ken Thompson's wife took their son on vacation. He had 3 weeks alone.
He wrote the kernel in week 1, the shell in week 2, the editor in week 3.
Modern dev teams spend 3 weeks in sprint planning.
Texas to Mangalore is about 1100-1200 nautical miles. Even if the tanker made 14 knots speed throughout the voyage, still she will take 35-36 days to complete the voyage. This means this tanker had started its voyage even before the war had started.
I honestly don’t know what people are on about but any work where you let the agent going on and on …
- beyond 25-30 turns
- beyond 50% context fill
- beyond 5-6 spawned sub agents
… is going to come out not just terribly shit, but also you’ll be unable to fix that output
I’ve started career in a startup which was an official ibm partner. I’ve little bit exposure into ibm I series , rpgle , db2 and COBOL.
The math looks good on paper, but $2/line code translation ignores the 'Engineering Gap.' You can't just move a 30-year-old RPG/COBOL system to Postgres on Cloud and expect the same results. AWS RDS gives you a 99.99% SLA (52 mins downtime/yr), but these AS/400 'tanks' regularly hit 10 years of continuous uptime.
Those who trolls has no idea about durgasoft. Anyone who was into job search during 2010 - 2015 would know who they are, similarly a lot of others used to be there like, jspiders, jdot..etc
For those who are wondering how to find this out, or how i find it out
1. Go until placing an order.
2. Inspect the console for the api call. You will see this is being triggered - airportfits.com/cgi-bin/api.…
3. If you do a curl on this : curl site.com/cgi-bin/api.py → server returned raw Python source instead of running it
4. Read the source → found DB_PATH = '../airportfits.db'
5. Translated to a URL → site.com/airportfits.db
curl -O site.com/airportfits.db → HTTP 200, file downloaded
6. sqlite3 airportfits.db "SELECT * FROM orders;" → 47 orders with full customer PII
AI can write the code, but it can't feel the friction. Software remains a human utility, and only humans have the lived experience to define the problems worth solving