pycloud.ai

Cloud and AI company building for today and the future

[pycloud.ai @pycloudsolution]

@yuzu_jpg @X +1

[pycloud.ai @pycloudsolution]

SWE is an art - products show the engineering minds that build them

[Sam (Knicks in 5) @futurenomics]

It’s wild how differently Codex and Claude Code feel.

Codex is a Stanford engineer: you give it the shape of an idea and it comes back in 5 minutes with a working prototype but with code that feels like it was written 5 red bulls deep at 4 am.

Claude Code is spiritually Eastern European: demands an annoying amount of clarification, goes on vacation for a couple of months, but then delivers something elegant enough to make Knuth cry.

[pycloud.ai @pycloudsolution]

And storage

[Bola Banjo @bolaabanjo]

Everything is compute.

[pycloud.ai @pycloudsolution]

"As the child delights in his mud pie, so the adult enjoys building things, especially things of his own design"

[pycloud.ai @pycloudsolution]

An ounce of prevention is worth a pound of cure - Benjamin Franklin

[pycloud.ai @pycloudsolution]

@heyblake https://t.co/cvenGFx4M1

[pycloud.ai @pycloudsolution]

Good read - covers fundamentals of SWE

[Dr Milan Milanović @milan_milanovic]

𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 ≠ 𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴

You shipped the feature, all tests passed, and the PR is merged. But who's maintaining that code in three years when you've moved on?

In 2005, over 80% of Google Search infrastructure pushes caused user-facing bugs. The fix wasn't better engineers. It was shifting testing left, meaning catching problems in the editor, not in production. Bugs dropped by half within a year.

Titus Winters put it simply: "Software engineering is programming integrated over time." Every function you write has a lifespan, a maintenance cost, and someone who'll inherit it without context.

The book "Software Engineering at Google" lays out principles I keep coming back to after 23 years of building software. Hyrum's Law says every observable behavior of your system becomes someone's dependency, documented or not. The Beyoncé Rule says if you liked that behavior, you should have put a test on it.

Google's engineers largely stopped using mocking frameworks because mocks test how something was done, not what actually happened. And DORA data keeps proving what we already know: small, frequent releases beat mega-deploys every time.

The teams that ship well aren't the ones with the best coders. They're the ones with the best systems around their code. I've seen this managing teams of 40+ engineers: process beats talent when talent doesn't have process.

The book is free at https://t.co/IHaWmQygFb. I broke down all 11 lessons with examples beyond Google.

👉 Full breakdown here: https://t.co/VimzA4ufhb

--
🎁 This newsletter is brought to you by @QodoAI: https://t.co/GhTaWuumN8

[pycloud.ai @pycloudsolution]

Energy can neither be created nor destroyed, it can only be transformed from one form to another

[Elon Musk @elonmusk]

@beffjezos Energy is the true currency

[pycloud.ai @pycloudsolution]

All men have fears, but the brave put down their fears and go forward, sometimes to death, but always to victory.

[pycloud.ai @pycloudsolution]

Hilarious 😂😂

[Jon C. Phillips @joncphillips]

Building a SaaS is basically one long decision tree:

• come up with the idea
• sanity-check if anyone actually needs it
• name it (then rename it three times)
• buy the domain
• design a logo at 2am
• scope the MVP
• cut half of it
• pick a stack
• second-guess that stack
• set up your repo, branches, CI
• design the database
• build the first models
• set up auth
• debug auth
• add OAuth because fuck it
• realize your data model’s wrong
• refactor everything
• pick a rich text editor
• hate all of them equally
• design the dashboard
• add 2FA (and lock yourself out of course)
• integrate payments
• add yearly plans
• handle upgrades, downgrades, and proration math
• write onboarding emails
• design success, warning, and error toasts
• design empty states
• build an async queue
• handle time zones (ugh…)
• setup background jobs
• forget to handle retries
• add logging
• fix logging
• add metrics
• fix metrics
• add a feature flag system
• create docs
• write transactional emails
• design error pages
• implement rate limiting (probably too late)
• build a settings page
• make a changelog
• update the roadmap (maybe)
• add dark mode
• break production
• fix it, add monitoring
• refactor the UI again because it looks dated already
• add analytics, events, metadata
• write privacy policy and terms
• ship
• immediately find a bug

That’s the real roadmap. And you’re doing all this while trying to figure out distribution.

[pycloud.ai @pycloudsolution]

If you can imagine it, you can build it. #MicrosoftAITour

[pycloud.ai @pycloudsolution]

Ship fast with https://t.co/AoLVE2Vx4r

[pycloud.ai @pycloudsolution]

Bend the curve of innovation
- Build with loveable
- CICD with Github
- Deploy with AWS Amplify

Impossible does not exist

[pycloud.ai @pycloudsolution]

@Microsoft Researcher Agent is fire 🔥

[pycloud.ai @pycloudsolution]

Implement Idempotency

[Ashutosh Maheshwari @asmah2107]

Quick question: Your order flow involves 5 microservices (Payment, Inventory, Shipping, etc.).

The Payment service succeeds, but the Inventory service fails.

You've taken money for a product you can't ship.

How do you manage this workflow to ensure data consistency across services?

[pycloud.ai @pycloudsolution]

I'd keep it simple and go with Azure web apps

[GPS @madebygps]

Want a project that will really test your skills? Try building something that deploys and works resiliently on Azure Spot VMs.

[pycloud.ai @pycloudsolution]

Solved a similar situation for a customer on AWS with SQS FIFO to achieve indempotency. Step functions helps as well.

[Ashutosh Maheshwari @asmah2107]

Your financial pipeline uses Kafka to process payments.

A worker sends the money, but crashes before committing the offset.

A new worker starts and processes the same payment again.

You've just double paid someone.

How do you fix this ?

[pycloud.ai @pycloudsolution]

pip install stripe

[pycloud.ai @pycloudsolution]

We help optimize for high availability and scalability to mitigate these risks and enhance customer experience

[‘Tunde Omotoye @TundeTASH]

There is a major hit on the IRCC server right now because of the thousands of people trying to submit their HCWIP application at the same time.

System keeps saying error for some people.