AI stuff,Kotlin,Flutter,Dart,CFML,Java,Linux and more,✍️@ix,💑 @blauerpunto,✈️ pilot,🎮 Nintendo 385233198305,🎙 @codecafeteria, tshirts&hoodies only. 🌈❤️💜💙 he/him
Wellington (Aotearoa - NZ)
Joined April 2007
- Tweets58.8K
- Following1.2K
- Followers2.2K
- Likes12.2K
Kai Koenig retweeted
Raygun4Flutter 3.7.1 is out.
A maintenance release with updated build and lint tooling, Dart 3.9 compatibility, and refreshed Flutter 3.47.2 lockfile metadata.
Requires Dart 3.9+ and Flutter 3.35+ raygun4flutter: ^3.7.1 pub.dev/packages/raygun4flut…
🤖 Made with AI
Kai Koenig retweeted
I sympathize with programmers who worked on systems they never cared about, for customers they didn't connect with, and took all their working joy from the mechanical act of turning tickets into code. You have to care about the software to be excited about making it w/ agents!
Kai Koenig retweeted
OpenAI have proven that matrix multiplication can be done in O(n^2.25) operations, which is a massive improvement. It is especially concerning because these AIs themselves run on matrix algebra.
The previous record was O(n^2.37), which had barely moved from O(n^2.4) in 1990.
OpenAI blew it away.
But... it is a nonconstructive proof. No algorithm was found, only a proof that it must exist.
Kai Koenig retweeted
An hour a day of maths was a key recommendation of the 2021 Royal Society Maths Report that was commissioned BY HIPKINS. He sat on that report for 2 years and did nothing. We implemented an hour a day in our first 100 days and used the rest of the recommendations as a blueprint for our 2024 Maths Action Plan. We now have 6% more students at curriculum expectations in maths at year 6. Hipkins sat and watched maths results tank while ignoring the answers right in front of him and he is still refusing to keep an hour a day in tonight’s debate. National listened to the experts and early results are improving for the first time in many years as a result. Our kids deserve a world-leading education and National are delivering it.
Kai Koenig retweeted
Most AI assistants look like a blob with two dots on them.
Snore.
The Autohive.com Bees take it to the next level 🫡
Kai Koenig retweeted
Remember when you had to pretend your browser was Internet Explorer?
Well welcome back…
Figma is desperately trying to be irrelevant.
Hey David, Gayani here from Figma. You're right that our remote MCP server only accepts clients on our supported list, and Pi isn't on it yet. You can see the current list in our MCP catalog at figma.com/mcp-catalog. If you'd like Pi considered for a future addition, please fill out this form: forms.gle/qSvUawwznWyoj8Go7.
Kai Koenig retweeted
i'm an airline pilot. i built an app with AI.
i'm not a software engineer. i'm not pretending to be.
but "you're not a real developer" misses the point.
i wasn't applying for the job.
i wanted a tool that worked the way i needed. not a second career.
Kai Koenig retweeted
For engineers having trouble with @dhh’s talk here some obvious truths about the future of programming:
You will not write code
You will not read code
You will not choose a language
You will design the architecture
You will define outcomes
You will build auto-delivery loops
Kai Koenig retweeted
Do yourself a favor and follow engineering leaders like @obie and @GeoffreyHuntley that can see the future of engineering without thinking it means the end of engineering
Kai Koenig retweeted
speculation to what happened with openai and medicare. specifically using sources from way back machine to look at server and application configuration…
Kai Koenig retweeted
🚨 Raygun4Android 6.2.3: important bugfix release
If you rely on Raygun to catch Android crashes, update as soon as you can.
This release fixes reliability issues that could lose crash data or, in one case, allow the crash reporter to crash the host app.
What changed?
🛟 Uncaught crash reports now persist to disk before process termination and deliver after the app restarts.
🔁 Pending reports stay retained until delivery reaches a final result, including across transient send, storage, or WorkManager failures.
🗑️ The on-device cache now evicts the oldest report at capacity and trims the oldest reports when you lower the storage limit.
🛡️ A failure in onBeforeSend or persistence can no longer crash the host app.
📦 Transitive OkHttp dependencies now align with the OkHttp 5.4.0 BOM, preventing mixed OkHttp module versions.
⚠️ Compatibility note: In 6.2.2 and earlier, setting maxReportsStoredOnDevice to 0 disabled offline caching. From 6.2.3, 0 is ignored and your current limit stays in place. Set a value from 1 to 64 instead.
There are no public API or ABI changes. The minimum SDK remains 23, and the compile SDK remains 36.
Get it now:
implementation("com.raygun:raygun4android:6.2.3")
Full changelog: github.com/MindscapeHQ/raygu…
A crash reporter needs to preserve crash data and stay out of your app's way. This release fixes both issues.
#Raygun #Android #AndroidDev #MobileDev #CrashReporting
🤖 Made with AI
Kai Koenig retweeted
"Why would I use Autohive instead of Claude?"
It shouldn't have to be a feature but "model routing so your business doesn't stop when Anthropic can't keep it up".
Kai Koenig retweeted
lol, to hell with all of this. I read the code, fix bugs, discuss the intent docs/PRD's, and communicate with my brain.
1. it's fun
2. the juniors are watching
This is an insane phase like OOP in the 2000's, The Internet years later, "mobile revolution" years later, Web 2.0 & "api all the things" years later, blockchain/NFT years later, tech over-hiring years later.
That said, for those of you who cannot hunker down and do things the right way "because management told us to do this insanity", I'm sorry. Hang in there and either weather it and then find another gig.
I am done with this shit. It is over. The state of engineering right now is horrible. It has been half a month since I started a new role at a big company. Nobody knows anything here. The specs, code, tests, PRDs, tickets, resolution of those tickets, reports, etc., everything is made by Claude Code. Nobody on my team likes this. They are being forced to ship as much as they can. I have heard multiple times from higher management that pushing code is not a bottleneck, so why are we slow? People are working 12 to 13 hours a day just to press enter. Nobody is reading anything. Humans in corporate are doing nothing on their own. Everyone, literally everyone, from an L1 to an L7 engineer here is doing the same thing. Talk to Claude. There is no sense of victory. Nobody is resolving bugs. In reality, nobody is thinking anymore. Everything is done by LLMs. It is so soul-sucking. I would not mind it, to be honest, if we were at least given the time to check out the code and see what is going where. But no, the goal is to just ship. No matter what happens.
Kai Koenig retweeted
"AI writes terrible code."
Does it work?
Yes.
Do the tests pass?
Yes.
Can another AI understand and modify it?
Yes.
Then I have terrible news about how much your definition of terrible code matter. Humans aren't reading this stuff for much longer, many already stopped.
Kai Koenig retweeted
Autohive Mobile 1.5.0 is now live on iOS and Android.
This release makes chat feel cleaner and more reliable, with a faster composer, in-chat PDF previews, and better history for meeting notetaker conversations.
What’s new:
+ Redesigned single-row chat composer
+ PDF previews directly in chat
+ Actioned notifications now clear automatically on Android
+ Meeting notetaker conversations now appear in chat history
+ Improved session reliability and chat participant handling
Kai Koenig retweeted
Hello, I’m the CEO of a AI lab that isn’t profitable & is burning investor capital at a significant rate.
Token for token I can’t compete against open source models.
So let’s try use over regulation to kill my competition & save my company’s business model.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Kai Koenig retweeted
someone wake @MikeJudge the fk up and make the next season of Silicon Valley already.
So much content left on the table with nowhere to mock it.
Kai Koenig retweeted
Maybe I’m misreading this but it shows it changing outputs for EU regulatory purposes?
That to me says:
1. We’re not paying for the best token anymore.
2. It makes Anthropic models questionable in fields where words matter, which is many fields.
