Also where is this idea that it takes days to ship Python in CDK coming from?
Edit: great, getting downvoted for daring to ask the author to explain a claim. I’m noticing other people asking questions getting downvoted, too. Brigading isn’t a good look.
Just to make it explicit, I (author) did not downvoted you.
Both me and sebst (co-author) tried to answer everybody and did so politely.
We respect other opinions and Stelvio (as any other thing) might and is not for everybody.
Regarding your question. Well, it depends on the size of the project. Small enough you can do anything fast. I'm working on bigger mostly serverless system on AWS and it did take lot of time to to setup everything with CDK, certainly days.
I believe it would be faster with stelvio as it offers higher level abstractions than CDK - Stelvio was born out of my frustraction with CDK, I was doing lot of things again and again, waiting long for deployments etc.
If someone is happy with CDK, then they should use CDK, it's good too, even for me much better than CF or Terraform.
But if you feel things could be even less cumbersome, with less code and faster, maybe you could give Stelvio a try.
We're not claiming we'll do everything for everybody, we have our opinions and Stelvio is opinionated - shamefully focusing on (app) developers rather than on infrastracture focused people. https://stelvio.dev/blog/why-i-am-building-stelvio/
Thank you for your comment and wish you a good day.
Stelvio's main selling point here is that you can use our higher-level components for different services and have them automatically configured.
So, you don't have to configure IAM roles, or Env vars manually, as this is handled for you through a concept called linking. https://stelvio.dev/concepts/linking/
In our experience, that alone adds a lot of productivity gains for teams.
Higher level CDK constructs do the same thing. But honestly, IAC is one of the easiest thing for LLMs to do and there is plenty of documentation to troubleshoot. There is no reason to introduce this into a company instead of using the official CDK.
But coincidentally Stelvio was born out of frustration with CDK which I'm using at my day job for 4 years at this point:
- slow deployment: CDK is layer on top of cloud formation, it first translates to CF which is then moved to AWS and resolved/deployed there. Process is quite slow and if something goes wrong it's hard to debug, rollbacks take ages, sometimes they block due to inter-stack dependencies
- CDK is still quite low level and focused on infra. You just can't create say api gateway with 3 routes each using 3 different lambdas with permission to use dynamo table in 4 lines - you an with stelvio
- whatever code change you need to test you need to deploy it first which is probably slowest with CDK(compared e.g. to pulumi) then even if you run it you can't really debug it or just see prints, you need to just go thru cloudwath or other services - stelvio allows you to run lambdas in "dev mode" so you don't need to redeploy and run your lambdas locally for instant feedback and even debugging support
Having said that CDK is good tool and I'm happy that it exists as I like it much better than CF itself or Terraform. Stelvio just tries to be even better and focused on developers.
Regarding LLMs sure, problem with LLMs is not they can't generate the code but if you're willing to read and understand all of it. Stelvio is less code with higher abstractions so it's easier to comprehend.
And what assurances are there that you will be around for five years? Or that you will support new services features when they come out?
And I always “disable rollbacks” this has been a feature in CloudFormation, CDK and SAM for years.
Running lambdas locally with SAM has been a feature for at least 5 or six years as with the CDK. But these days you really should be packaging lambdas as Docker containers - those are really easy to test locally without any special infrastructure
We still believe to have a more flexible solution that also adds some features, including combining multiple cloud providers which at the moment we use to enable cloudflare DNS in front of AWS infra. Feel free to give it a try!
I still use my dad’s old HP 15C. Form factor is good, aesthetically it’s very appealing, it feels efficient to use, I like the tactile feel of the buttons, and I like thinking of my dad when I use it.
> Is this a flaw in the cryptography itself?
No. The underlying cryptographic algorithms (3DES and AES-128) remain secure. The vulnerabilities arise from:
Protocol design choices that allow unauthenticated memory writes after initial authentication
Lack of atomicity when writing cryptographic keys across multiple memory pages
Widespread misconfiguration in real-world deployments (unlocked memory, static keys)
Non-NXP compatible chips with severely flawed random number generators
I’d always hated running. I’ve done C25K a few times with no problem, but once I reached that goal I felt I couldn’t make much additional progress. But the need to move and running’s simplicity and minimal requirements meant I kept trying every couple of years. I read/watched a little and got some tips from ultra runners and actually started to make progress. I was actually enjoying it and looking forward to running. Then I got the worst fucking shin splints imaginable where I had to nearly crawl halfway back home. Stopped running to heal, which took a very long time. I’ve tested a few runs but there’s still pain and I just don’t want to go through that again.
So now I row every day. I get a much better exercise high from rowing, progress is much more noticeable, it’s improved vitals more than running has, there’s no pain, and I don’t have to worry about weather. I occasionally miss the change of scenery or things like running on a cold snowy day but I can just go rent some cross country skis when I get that itch.
I had a similar experience with running, including terrible shin splints that took me out for weeks at a time.
I went to a "run clinic" where they observed my gait. I'm paraphrasing here since this was many years ago, but basically they said that my stride was slightly too large and that my knees were behind my feet during the foot strike. My cadence was around 150-155 steps per minute and they suggested increasing it to 170-180, basically meaning my steps would be smaller but more frequent.
I downloaded a metronome app on my phone and set it to 172 to make sure that I maintained the proper rhythm while running. Worked immediately and I never had shin splints again.
I’ve long felt estimations are a negotiation. It’s not as much about how long it will take or how much it will cost, but what do you really need and what can you afford. Kinda like I’m helping someone buy a car based on their needs and budget, I’ll put together 3 “trim line” options:
1. Economy (bare functionality, prefer low cost and fast delivery over reliability and longevity)
2. Mid tier (good enough quality and reliable but no frills)
3. Luxury (all the bells and whistles)
The business will want all the bells and whistles of 3 but also the pragmatism of 2 but the budget and timeline of 1. So, they don’t actually pick themselves, I choose for them based on the circumstances and we negotiate the finer points.
Devs can gold plate the shit out of any project, but having a plan for #1 is helpful when shit hits the fan and the business needs you to pivot. More than that, it’s useful during the negotiation to illustrate the ramifications of shortcuts. It’s helped me more than a few times to avoid some really stupid decisions from panicking PMs and execs.
I think in 2026 it's hard to make a video look this "bad" without it being a clear aesthetic choice, so not sure you could find this video in another setting.
I really disagree with the label brainrot. Brainrot is low-quality garbage with no artistic merit, and very little thought behind its creation, which does nothing but make you briefly pause while scrolling, before scrolling away with no lasting impression being done to your mind (besides increased boredom and inability to focus).
This is clearly an artistic statement, whether you like the art or not. A ton of thought and time was put into it. And people will likely be thinking and discussing this video for some time to come.
I think I have a different definition of brainrot than you. I think of attention grabbing visuals that have no meaning, just something to keep your eyes glued to the screen.
A good example are those Subway Surfers split screen videos, where someone is babbling about nothing in one frame, but the visuals in the other keep people watching.
Another example is AI-narrated “news” on YouTube. Nobody would normally listen to an AI voice read AI slop, but if there are some extreme video clips quickly switching every few seconds, people don’t immediately click away.
Brain rot shreds the attention span and uses all kinds of psychological tricks to keep people engaged. In the Helicopter video, every second is packed with visual information, not to contribute to the narrative but to capture attention. The backgrounds are full of details. The camera never stops moving. The subjects depicted are even attention grabbing: police lights, dancing people, guns, car crashes, flamethrowers! Hey, does that guy have pink curlers in his hair?
It’s not that I don’t like it (I kinda do), but a media diet of that kind of content is bad for the brain.
> Another interpretation would be to conclude cynically that it’s time humans get either enhanced or replaced by a more powerful form of intelligence.
Perhaps the author is just gaming out a thought experiment, but I’ll just take it at face value. I am genuinely baffled by the obsequiousness some people display regarding LLMs. Let’s assume it really is a more powerful form of intelligence (ugh) and it “replaces” people, how do you think that ends for you?
You are trying to convince yourself that you’ve happened upon a benevolent god that truly, deeply understands you while staring into a reflection pool.
Mockito was indeed a poor fit for Kotlin. MockK is the one. Except I suppose for shops that have projects that mix Java and Kotlin and already have a Mockito tests.
As a reader interested in Framework and looking for honest feedback in the wild, this was a really frustrating blog entry. It feels like at every turn you’re coming up with conflicting requirements and impossible demands that even your X1 doesn’t meet. For example: regarding the MacBook, the worry about long term support for Asahi is understandable, but one of your main criticisms is that it has the same battery life as other Linux laptops? You wanted a bigger, repairable laptop and didn’t expect it to be larger or heavier? You preferred the slightly more washed out red of your old laptop, but didn’t tell us whether you could fix it with a display setting or why having more vibrant colors or having to adjust display settings is a bad thing for your use case? Have you considered just turning on a light rather than working in a dark room (please turn on a light, I destroyed my eyesight working like that)? I’m devastated to read you found the battery life simply par and that it used a little bit of electricity after suspend for 8 hours like every other laptop on the market, and that you were able to turn the power led off but not in your preferred location.
At the end of the day, you’re not required to like a thing for any reason you see fit, but I very much hope you figure out what you actually need and that such a device actually exists on this planet before you purchase another one.
Edit: great, getting downvoted for daring to ask the author to explain a claim. I’m noticing other people asking questions getting downvoted, too. Brigading isn’t a good look.
reply