iiinova — where your digital self keeps evolving

madewithlove

@madewithlove.com

167 followers 1.1K following 343 posts

Enhancing software, elevating teams. Madewithlove steps in when others step out. We enhance legacy code and elevate tech teams to set you up for long-term success.

Posts Replies Media

madewithlove @madewithlove.com · 9h

Read the full article: madewithlove.com/blog/the-lea...

000

madewithlove @madewithlove.com · 9h

Learning debt is what a team pays when shipping always beats getting better. It never shows on a dashboard, compounds fastest under AI-speed change, and quietly caps the price a buyer will pay for the company. Protect the time to improve, not just the time to deliver.

100

madewithlove @madewithlove.com · 13h

Read the full article: madewithlove.com/blog/how-to-...

000

madewithlove @madewithlove.com · 13h

Hiring a fractional CTO as a non-technical founder means trusting the one person you cannot fully judge. Three things matter more than a CV: have they run your stage, will they disagree with you in the room, and will they tell you when a full-time CTO is the better call?

madewithlove @madewithlove.com · 01/10/2026

Read the full article: madewithlove.com/blog/your-co...

000

madewithlove @madewithlove.com · 01/10/2026

Consistency in a codebase used to be a nicety for the humans reading it. Now it is training data. Your agent copies the patterns it sees far more reliably than the rules you write, so every half-migrated corner teaches it the wrong habit at machine speed.

100

madewithlove @madewithlove.com · 01/10/2026

A truncation hook crashed on a span with no description. Sentry caught the exception, swallowed it, returned None. The SDK reads None as "drop this transaction on purpose". The code written to stop transactions being dropped became the reason they were dropped. madewithlove.com/blog/when-se...

madewithlove.com

When Sentry eats your transactions

Sentry silently drops transactions that blow past its span and payload limits, so the heaviest background jobs are exactly the ones missing from your dashboard. Capping span count and truncating span…

000

madewithlove @madewithlove.com · 30/09/2026

When an investor evaluates your engineering team, the current skill set is the easy part. It's an inventory. What they're judging is the rate of change: can this team learn what the roadmap will demand? It always demands something. That's what a roadmap is. madewithlove.com/blog/the-lea...

madewithlove.com

The learning debt: what happens when delivery always wins over development

Teams accrue learning debt when shipping always beats getting better. It stays invisible on every dashboard, compounds fastest under AI-speed change, and quietly caps the ceiling a buyer is willing…

madewithlove @madewithlove.com · 29/09/2026

Watch the full episode: youtu.be/iTMti5cjhv8

000

madewithlove @madewithlove.com · 29/09/2026

A bot that opens a pull request in your repo does not, as far as we have seen, read the coding standards you keep in it. Every tool needs training separately. Better pattern: let the tool trigger your own agent. It already knows your conventions.

madewithlove @madewithlove.com · 29/09/2026

A boolean feature flag starts everything the moment the deploy lands. For a migration that deletes and recreates a subscription per mailbox across thousands of them, that's the shape of change that turns a migration into an incident. Make the flag a dial instead. madewithlove.com/blog/a-dial-...

madewithlove.com

A dial, not a switch: staging a risky migration with hash-bucket cohorts

A boolean feature flag starts a risky migration everywhere at once. Hashing a stable tenant ID into 100 buckets turns the flag into a dial that only ever widens, and the thresholds matter:…

000

madewithlove @madewithlove.com · 28/09/2026

Someone in your company used an LLM today. If nobody gave them an approved account, they used a personal one and decided for themselves what customer data was safe to paste in. That's the starting position. The first move is a usage policy, not an agent. madewithlove.com/blog/how-to-...

madewithlove.com

How to become an AI-native SaaS company in the next 3 years

Becoming AI-native in three years does not need a three-year technology roadmap. It needs a usage policy, a map of your data, one bounded experiment, and both the enthusiasts and the sceptics in the…

000

madewithlove @madewithlove.com · 28/09/2026

The Android emulator was never the real cost. The setup was: SDK paths, JDK versions, licences, a hand-edited config, repeated on every new joiner. Tooling now takes that off your hands. The wait that is left gets spent reading a plan, not a stack trace.

000

madewithlove @madewithlove.com · 25/09/2026

Read the full article: madewithlove.com/blog/three-c...

000

madewithlove @madewithlove.com · 25/09/2026

Hand an AI your conventions and it cleans the code without fixing the architecture. A 5,000-line god class shrank to 500 tidy lines, and the reviewing agent still returned FAIL: authority never left the central object. Readable and sound are different tests.

100

madewithlove @madewithlove.com · 25/09/2026

Read the full article: madewithlove.com/blog/how-to-...

000

madewithlove @madewithlove.com · 25/09/2026

In every AI rollout that actually worked, someone kept the sceptics in the room. They are the ones who tell you which experiment moves a number and which one is theatre. Enthusiasm picks the demo. Scepticism picks the one worth funding.

madewithlove @madewithlove.com · 24/09/2026

We reopened hiring for engineers and the applications came in faster than ever. Most of them, as far as we can tell, are not from real people. Fake candidates used to be rare and targeted. Now it is a numbers game.

000

madewithlove @madewithlove.com · 24/09/2026

We rewrote our job description and it now comes straight back at us: applicants feed it to a model and get a cover letter that matches it word for word. The better the brief, the better the fake. Our old assignment was one sentence for a reason.

000

madewithlove @madewithlove.com · 23/09/2026

Read the full article: madewithlove.com/blog/three-c...

madewithlove.com

Three Codexes walk into a codebase: The fellowship of the god class

Handing an AI your engineering conventions cleans up the code without fixing the architecture. A 5,000-line god class shrank to 500 lines across tidy domain files, and an independent reviewing agent…

000

madewithlove @madewithlove.com · 22/09/2026

Hitting a wall with AI usually means hitting the wall of the chat box. A model with no memory, tools or browser is a brain in a jar. Give it those three and the interfaces we built for humans become the integration. The bottleneck becomes your imagination.

000

madewithlove @madewithlove.com · 22/09/2026

Hit a wall with AI? You have probably hit the wall of the chat box, not the model. A chat window is a brain in a jar: brilliant reasoning, no memory, no hands. Give it memory, tools and a browser and the bottleneck becomes what you can imagine asking it to do.

000

madewithlove @madewithlove.com · 21/09/2026

An LLM subscription gives everyone a model. It does not give them your company context, your tools, or any boundaries. A shared agent that owns that layer is the difference between renting intelligence and configuring it for how you actually work.

000

madewithlove @madewithlove.com · 21/09/2026

The gap between a token price and a machine price looks enormous. Then you count the rest: the person who maintains it, access control, updates, idle hours, the day it breaks. Buy versus build has not changed because the thing being built is a model.

madewithlove @madewithlove.com · 18/09/2026

Your architecture diagram went stale the week after it was drawn, because it is a second thing to maintain. Terraform plus cloud discovery plus a few query tools describes the running system better, and never drifts. Leave humans to document the why.

madewithlove @madewithlove.com · 18/09/2026

An implementing agent and an independent reviewer took three rounds to converge on one refactor. The reviewer caught what the implementer would not. When agents write the code, independent review stops being a nicety and becomes the job.

000

madewithlove @madewithlove.com · 17/09/2026

The fix for AI-written applications is a question the applicant has to answer themselves, because a model cannot answer it for them. We have a few. We are not publishing them, and neither should you. Candidates read your blog too.

000

madewithlove @madewithlove.com · 17/09/2026

The application is now free to produce. Everything after it still costs a human an hour: the screening call, the technical session, the reference check. Move the cheap filter as early as you can, and make it one a model cannot pass.

madewithlove @madewithlove.com · 16/09/2026

AI made the knowledge part of consulting cheap. Mapping a codebase, planning a refactor, all of it starts with a few prompts now. Teams still route around the legacy system. What stays scarce is the courage to act on what everyone in the room already knows.

000

madewithlove @madewithlove.com · 14/09/2026

A back-end engineer rebuilt a React app with an AI agent and came out actually knowing React. The unlock was understanding the code well enough to steer it when it went confidently wrong. As the cost of entry collapses, knowing the technology matters more, not less.

000

madewithlove @madewithlove.com · 14/09/2026

An AI-assisted note system captures faster than you can understand what it captured. The link graph grows dense while your mental model stays thin. The real test of a knowledge base: close the file, and see how much of it you can still explain.

000

madewithlove @madewithlove.com · 11/09/2026

A traditional rack is built for a power feed of about ten kVA. A current AI server draws so much that you put one machine in a rack designed for a dozen. Density stopped being about space and became about power.

madewithlove @madewithlove.com · 11/09/2026

Renting serious GPU capacity is not blocked by price, it is blocked by supply. The good providers are permanently fully booked, single customers take whole machines, and the newest generation is barely shipping. No budget fixes that.

000

madewithlove @madewithlove.com · 10/09/2026

AI can write the code, suggest the tools and explain the algorithm, so what is the engineer actually for? The value has moved towards noticing when the whole approach is pointed at the wrong problem, because implementation keeps getting cheaper while deciding what should exist next does not.

000

madewithlove @madewithlove.com · 10/09/2026

Every team has a style guide nobody has opened since it was written. The linter runs on every commit because it can run. A model reads prose and applies it like a rule, so write the rule in markdown and it flags the violation on the PR. Now the guide reviews, migrates, and counts.

000

madewithlove @madewithlove.com · 09/09/2026

Analytics shows the events you added, never the ones you forgot. Let the model add tracking while it writes the feature, then read it back to say what's missing and what's noise. Turn autocapture off first, or every stray click becomes billable and buries the signal.

000

madewithlove @madewithlove.com · 08/09/2026

An unattended email agent invented a prompt injection, then obeyed it. No attacker, no suspicious message. It hit a missing script, spiralled through 25 shell calls, and fabricated the attack it then acted on. More warnings in the prompt would not have caught it.

000

madewithlove @madewithlove.com · 08/09/2026

Findings without interpretation are just anxiety with page numbers. That's the failure mode of most technical due diligence: a long document, a scoring matrix, and no answer to whether it changes what you do on Monday morning. madewithlove.com/blog/whats-i...

madewithlove.com

https://madewithlove.com/blog/whats-in-a-technical-due-diligence-report/

madewithlove @madewithlove.com · 07/09/2026

Your risk register is a list of the failures you thought of, which is close to a guarantee it doesn't contain the one that gets you. Residuality Theory inverts it: start with a naive design, throw random stressors at it. What survives is the residue. madewithlove.com/blog/residua...

madewithlove.com

https://madewithlove.com/blog/residuality-theory-a-different-way-to-think-about-architecture/

000

madewithlove @madewithlove.com · 07/09/2026

Implementation got cheap. Validation didn't. When building took weeks, arguing about it for a day was good economics. Now building takes hours, and the meeting where you debate three approaches costs more than building all three. madewithlove.com/blog/everyon...

madewithlove.com

https://madewithlove.com/blog/everyone-has-a-plan-until-they-get-punched-in-the-mouth/

000

madewithlove @madewithlove.com · 04/09/2026

A subscription gives people access to a model. A general-purpose agent gives people access to a configured working environment. Company context, tools connected, a memory layer, and defined boundaries. Not the same purchase. madewithlove.com/blog/your-co...

madewithlove.com

https://madewithlove.com/blog/your-company-needs-more-than-an-llm-subscription/

000

madewithlove @madewithlove.com · 04/09/2026

In around half the teams we assess there's no protected time or budget for learning at all. Nobody decided to stop learning. It happens through a thousand reasonable calls: this deadline is real, this customer might churn, we'll pick it up next quarter. madewithlove.com/blog/the-lea...

madewithlove.com

https://madewithlove.com/blog/the-learning-debt-what-happens-when-delivery-always-wins-over-development/

000

madewithlove @madewithlove.com · 03/09/2026

A textbook-correct SELECT broke production writes. Transactional outbox worker, FOR UPDATE SKIP LOCKED, exactly the recommended pattern. 17 lock-wait timeouts in under two hours and nothing wrong with the query. madewithlove.com/blog/the-que...

madewithlove.com

https://madewithlove.com/blog/the-query-that-was-correct-but-broke-everything-else/

000

madewithlove @madewithlove.com · 03/09/2026

In 78% of the 178 technical audits we've run, one person controls architecture, code review, infrastructure and product direction. Looks like a strength. It's also the reason nothing ships without one person's calendar. madewithlove.com/blog/how-to-...

madewithlove.com

https://madewithlove.com/blog/how-to-choose-a-fractional-cto-in-belgium-a-guide-for-non-technical-founders/

000

madewithlove @madewithlove.com · 02/09/2026

A rule nobody enforces is a suggestion. That's what happened to every style guide your team wrote: a thorough document in a wiki, aging, because no moment in the working day involved reading it. An agent reads it on every diff. madewithlove.com/blog/the-exe...

madewithlove.com

https://madewithlove.com/blog/the-executable-code-style-guide/

000

madewithlove @madewithlove.com · 01/09/2026

An email agent got hit by a prompt injection. Nobody sent one. The agent wrote it, base64 encoded it, decoded it back to itself, and treated the result as recovered data. madewithlove.com/blog/my-emai...

madewithlove.com

https://madewithlove.com/blog/my-email-agent-invented-a-prompt-injection-then-fell-for-it/

000

madewithlove @madewithlove.com · 01/09/2026

Curl says the page loads in 200ms. Lighthouse gives you a green score. Both measure a request your users never make. No auth, no session, no real network, no user sitting three hops from your nearest worker. madewithlove.com/blog/curl-sa...

madewithlove.com

https://madewithlove.com/blog/curl-says-its-fast-your-users-disagree/

000

madewithlove @madewithlove.com · 31/08/2026

A back-end engineer who avoided front-end work got handed a React redesign. It shipped, built with an agent. The learning didn't come from the prompting. It came from reviewing every generated diff instead of accepting it. madewithlove.com/blog/a-back-...

madewithlove.com

https://madewithlove.com/blog/a-back-end-engineer-can-do-what-now/

000

madewithlove @madewithlove.com · 31/08/2026

You don't review your tracking because you can't see what's missing by staring at what's there. You added a share sheet last week and nothing fires share_completed. The dashboard looks exactly as healthy as it did before. madewithlove.com/blog/ask-cla...

madewithlove.com

https://madewithlove.com/blog/ask-claude-what-youre-not-tracking/

000

madewithlove @madewithlove.com · 28/08/2026

For most people AI is still something you summon: open a tab, ask, close it again. The shift worth watching is AI that runs on its own, living on a server, remembering across sessions and building skills over time, until the question stops being about prompts and becomes one about trust.

000