Andrii Reinvald

Senior Software Engineer at Klarna

[Vik @onlinedopamine]

...

[Marc Lou @marclou]

.@mntruell launched Cursor 8 times, and nobody cared.

Don't give up.

[Andrii Reinvald @andreinwald]

Hey @elonmusk when Grok on Apple Watch? Voice mode.

[Steren @steren]

Wow, it actually *is* DOM
(See me using Chrome DevTools to inspect the DOM)

HTML is <canvas> is going to open a new era of web UIs

[dex @dexhorthy]

computer, show me a perfect programming language

[can @can]

your db query is slow? just add this! boom, now you are ai-native instead of lazy!

[Jason Fried @jasonfried]

A bespoke software revolution? I don't buy it.

It'll exist. It already exists. Small consultants and big consulting firms have made custom software for years. It almost always sucks. It’s bloated, confusing, and because the client pays, it’s built wrong in all the ways.

Who’s excited about bespoke software? Software makers! Of course they're excited about building bespoke software — that's what they do. X is full of them. Your feed is full of people who love making software talking about making software. Of course they’re excited about the revolution. Echo, echo, echo...

Most people don’t like computers. Nobody in tech wants to say that out loud. People tolerate computers. They use them because they have to. Given the choice, most would rather not think about them at all.

So when someone suggests that AI means everyone will build their own custom tools, ask who "everyone" is. The three-person accounting firm drowning in client paperwork? They want the paperwork gone, not a new system to maintain. The regional logistics company with 40 trucks? They want the routes optimized, not Joe spouting off about this new system he’s been messing around with. The law firm billing 70-hour weeks? They want leverage on their time, not a software project to design.

They don’t hate technology. But building and maintaining their own critical systems isn’t their wheelhouse, regardless of how much faster and easier it’s become. It's another job on top of the job.

Will these people use AI? Absolutely, for all sorts of things. Will some outliers go deep and build real custom systems? Sure, but they're almost always people who already had some pull toward software. The curiosity was already there. They were dabblers before.

Giving everyone access to software building tools doesn't mean everyone becomes a builder. A powerful excavator doesn't turn a homeowner into a contractor. Most people just want the hole dug by someone else. They don’t want the responsibility either.

[Andrii Reinvald @andreinwald]

@ChromiumDev Another WebMCP demo on React
https://t.co/9SKaaHkN1j

[dax @thdxr]

everyone's talking about their teams like they were at the peak of efficiency and bottlenecked by ability to produce code

here's what things actually look like

- your org rarely has good ideas. ideas being expensive to implement was actually helping

- majority of workers have no reason to be super motivated, they want to do their 9-5 and get back to their life

- they're not using AI to be 10x more effective they're using it to churn out their tasks with less energy spend

- the 2 people on your team that actually tried are now flattened by the slop code everyone is producing, they will quit soon

- even when you produce work faster you're still bottlenecked by bureaucracy and the dozen other realities of shipping something real

- your CFO is like what do you mean each engineer now costs $2000 extra per month in LLM bills

[Harshit.exe @ishtwts]

Ask a programmer to review 10 lines of code, he’ll find 10 issues.

Ask him to review 500 lines,
he’ll say looks good 👍

[ThePrimeagen @ThePrimeagen]

so you are one of the 3

1. intentionally not trying to understand things
2. have no idea what 2010-2015 professional web 100% test coverage madness was like
3. something else that is neither ignorant or intentional misinterpreting

ill pretend 3 for kindness sake.

100% test coverage has and always will lead to a bunch of gamified tests that are breaking the blackbox testing contract and secretly testing implementation of single if statements somewhere deep within your code.

its a source of perennial problems. every change comes with many borked unit tests and it becomes very difficult to distinguished between broken because of internal implementation details change and broken because of some real bug

lastly, when you ask claude to do this, they will inevitably start adding testing data and leaking abstractions leading to significantly more brittle changes.

Yes, this is a bad idea

Please dont think that SOTA == savior

[Captain Eli @TheCaptainEli]

Grok’s Tasks feature is a game-changer!

Schedule any prompt to run automatically—daily trend reports, custom reminders, market updates, or whatever you need—delivered straight to your notifications. Smarter automation, powered by xAI.

Try it now!

[Abhishek Singh @0xlelouch_]

At a high level, every database has to answer the same question: when do I pay the cost of keeping data organized?

You can pay now, during the insert, or pay later, in the background.

MySQL (InnoDB) mostly pays up-front. Rows live inside a clustered B-tree, so every insert has to find the right place in the tree. That means pointer chasing, possible page splits, and immediate rebalancing. Inserts are more expensive, but once the row is written, reads are fast and predictable because the data is already well organized.

Postgres leans more toward deferring work. A new row is appended into a heap page, which makes inserts simpler and often faster. Indexes are updated separately, and space cleanup is pushed to VACUUM later. Writes feel cheaper in the moment, but the system pays over time through background maintenance.

The same tradeoff shows up with B-trees vs LSM trees.
B-trees do more work on every write to keep the structure sorted. LSM trees accept writes cheaply into memory and logs, then constantly merge and compact data later. Writes look fast, but you’re committing yourself to ongoing background work.

Logs (WAL, binlog) smooth this out by letting databases acknowledge writes before all disk structures are updated. But the work doesn’t disappear - it just moves to checkpoints, flushes, and compaction.

So essentially databases don’t avoid work. They just decide when to do it.
Fast inserts usually mean deferred cleanup. Clean layout usually means slower inserts. You’re always choosing where to pay.

[Peter Girnus 🦅 @gothburz]

Last quarter I rolled out Microsoft Copilot to 4,000 employees.

$30 per seat per month.

$1.4 million annually.

I called it "digital transformation."

The board loved that phrase.

They approved it in eleven minutes.

No one asked what it would actually do.

Including me.

I told everyone it would "10x productivity."

That's not a real number.

But it sounds like one.

HR asked how we'd measure the 10x.

I said we'd "leverage analytics dashboards."

They stopped asking.

Three months later I checked the usage reports.

47 people had opened it.

12 had used it more than once.

One of them was me.

I used it to summarize an email I could have read in 30 seconds.

It took 45 seconds.

Plus the time it took to fix the hallucinations.

But I called it a "pilot success."

Success means the pilot didn't visibly fail.

The CFO asked about ROI.

I showed him a graph.

The graph went up and to the right.

It measured "AI enablement."

I made that metric up.

He nodded approvingly.

We're "AI-enabled" now.

I don't know what that means.

But it's in our investor deck.

A senior developer asked why we didn't use Claude or ChatGPT.

I said we needed "enterprise-grade security."

He asked what that meant.

I said "compliance."

He asked which compliance.

I said "all of them."

He looked skeptical.

I scheduled him for a "career development conversation."

He stopped asking questions.

Microsoft sent a case study team.

They wanted to feature us as a success story.

I told them we "saved 40,000 hours."

I calculated that number by multiplying employees by a number I made up.

They didn't verify it.

They never do.

Now we're on Microsoft's website.

"Global enterprise achieves 40,000 hours of productivity gains with Copilot."

The CEO shared it on LinkedIn.

He got 3,000 likes.

He's never used Copilot.

None of the executives have.

We have an exemption.

"Strategic focus requires minimal digital distraction."

I wrote that policy.

The licenses renew next month.

I'm requesting an expansion.

5,000 more seats.

We haven't used the first 4,000.

But this time we'll "drive adoption."

Adoption means mandatory training.

Training means a 45-minute webinar no one watches.

But completion will be tracked.

Completion is a metric.

Metrics go in dashboards.

Dashboards go in board presentations.

Board presentations get me promoted.

I'll be SVP by Q3.

I still don't know what Copilot does.

But I know what it's for.

It's for showing we're "investing in AI."

Investment means spending.

Spending means commitment.

Commitment means we're serious about the future.

The future is whatever I say it is.

As long as the graph goes up and to the right.

[Sebastian Siemiatkowski @klarnaseb]

People think BNPL is “too good to be true” only because they don’t realize HOW INSANELY profitable U.S. credit cards are:

Merchants pay ~2%.
Revolving users add ~2.5%.

A $100 transaction actually costs $5.50 🤯

So yes you can offer something nicer and still make money...

[Andrii Reinvald @andreinwald]

@samlambert If someone not ready to pay 20-50$ for db where ALL data of their users - he doesn't care about them.

[Guillermo Rauch @rauchg]

Vercel started out similarly, but imagine these Nginx servers being all around the globe, and the metadata (hostnames, routing) being incredibly dynamic.

Each git push creates a new entry in this metaphorical “config”, and there’s a billion entries. The write throughput is constant, as people can roll back, push, and even change routes dynamically via our Firewall.

Because all these entries can’t possibly fit in one server, you have the problem of creating tiered caches that don’t introduce performance bottlenecks for old entries (otherwise you get “cold boots” of metadata itself).

An additional complexity is that the “origin database” of this system should be able to be offline. If a CDN region gets partitioned from the network, it should still be able to serve production traffic.

This metadata system has had to be re-architected 4 times as we’ve reached different levels of scale and also ambition. We actually set out to make the original design as monolithic as possible, but scale will push you, for good reason, into services (whether we can call them “micro” or “macro” is debatable).

If you’re interested in working on the infrastructure that accelerates the deployment rate of the world, my DMs are open!

[Andrii Reinvald @andreinwald]

@Klarna is officially public! 🎉 100M+ consumers, countless innovations and this is only the beginning. #Klarna #NYSE #IPO #Fintech #$KLAR

[@levelsio @levelsio]

I already tried ads repeatedly and ROI was always negative

I hired the best people in the industry and ROI was still negative

I think the reason is basic economics:

I compete with VC-funded companies with $100M balance sheets who happily enjoy negative ad ROI if it means growing their user base

They can pay $1 per ad, lose $0.10, make $0.90, and it's still good because it means GROWTH. And growth means being able to raise another $100M for the next round. Then at some point they try push to profitability and IPO!

I can't do that cause I'm self funded, so everything I do has to be profitable

If you're competing for the same ad slot with those companies, you can't economically have positive ROI for very long. For a while yes, not for very long.

[Andrii Reinvald @andreinwald]

@levelsio Cinema and TV often use a few cameras with different zoom levels or perspective because a static video plan is boring. Best would be rotating head to 45° with zoom and slight background shift (person looked into another camera)