Adrian Kajda retweeted
Remek is currently working on a decision model (i.e. "system one" model, similar to Jev / OpenAI's Decisions API) and he's currently ranked #1 in open weights (#2 overall!) on JevBench, essentially beating almost every commercially released product, including Jev itself that went viral some time ago
what's even crazier is that he's working on it alone, as an after-hours side project, he started a few (two?) weeks ago and he's paying for compute out of his own pocket
basal-omni 🇵🇱 - TOP1 🤩🤩🤩 🌍 na świecie w modelach otwartych oraz TOP2 wśród zamkniętych i otwartych
Wg niezależnego benchmarku modeli system-one gdzie testowane są wszystkjie modele na świecie - JevBench (link do bencha w pierwszym wpisie). Czyli można? Można! Polskie AI 🇵🇱
basal wyprzedził model OpenAI, Microsoft oraz pierwowzór z TypeSafe (jev).
Thank you, @airesearch12 , for the independent evaluation on JevBench.
Adrian Kajda retweeted
Mimo że dostawcy AI szyfrują procesy myślowe modeli i odsyłają je klientom do stanowej pamięci szyfrem, badacze wykazali, że metadane API pozwalają odtworzyć pełny chain-of-thought z dokładnością 1:1
Adrian Kajda retweeted
You can now use your ChatGPT subscription in the Zed agent, with the same usage and rate limits you benefit from in Codex directly. We're grateful that @openaidevs continues to support subscription-based access for third-party tools, even as others move toward usage-based billing.
Adrian Kajda retweeted
With the help of Claude Mythos Preview, the Firefox team fixed more security bugs in April than in the past 15 months combined.
Adrian Kajda retweeted
I upgraded my Claude token counter tool to compare different models and Opus 4.7 does appear to use 1.46x times the tokens for text and up to 3x the tokens for images - it's priced the same as Opus 4.6 on a per-token basis so this is actually a pretty big price bump
ALT Screenshot of a token comparison tool with an uploaded screenshot PNG image. Models to compare: claude-opus-4-7 (checked), claude-opus-4-6 (checked), claude-opus-4-5, claude-sonnet-4-6, claude-haiku-4-5. Note: "These models share the same tokenizer". Blue "Count Tokens" button. Results table — Model | Tokens | vs. lowest. claude-opus-4-7: 4,744 tokens, 3.01x (yellow badge). claude-opus-4-6: 1,578 tokens, 1.00x (green badge).
Adrian Kajda retweeted
Gemma 4 is now available in Android Studio!
By running Gemma 4 locally on your machine you get access to AI code-assistance that doesn't require an internet connection or an API key for its core operations — so you’ll never run out of token quota → goo.gle/4c1mDOD
Adrian Kajda retweeted
New supply chain attack this time for npm axios, the most popular HTTP client library with 300M weekly downloads.
Scanning my system I found a use imported from googleworkspace/cli from a few days ago when I was experimenting with gmail/gcal cli. The installed version (luckily) resolved to an unaffected 1.13.5, but the project dependency is not pinned, meaning that if I did this earlier today the code would have resolved to latest and I'd be pwned.
It's possible to personally defend against these to some extent with local settings e.g. release-age constraints, or containers or etc, but I think ultimately the defaults of package management projects (pip, npm etc) have to change so that a single infection (usually luckily fairly temporary in nature due to security scanning) does not spread through users at random and at scale via unpinned dependencies.
More comprehensive article:
stepsecurity.io/blog/axios-c…
🚨 CRITICAL: Active supply chain attack on axios -- one of npm's most depended-on packages.
The latest [email protected] now pulls in [email protected], a package that did not exist before today. This is a live compromise.
This is textbook supply chain installer malware. axios has 100M+ weekly downloads. Every npm install pulling the latest version is potentially compromised right now.
Socket AI analysis confirms this is malware. plain-crypto-js is an obfuscated dropper/loader that:
• Deobfuscates embedded payloads and operational strings at runtime
• Dynamically loads fs, os, and execSync to evade static analysis
• Executes decoded shell commands
• Stages and copies payload files into OS temp and Windows ProgramData directories
• Deletes and renames artifacts post-execution to destroy forensic evidence
If you use axios, pin your version immediately and audit your lockfiles. Do not upgrade.
Adrian Kajda retweeted
MCPs are the opposite of dead. They are the life blood of how AI agents use services inside mid-sized and above companies.
Case in point: Uber runs on MCPs internally, for good reason. Details: newsletter.pragmaticengineer…
Adrian Kajda retweeted
I made the hard decision to retire my pinned thread with programming tips. I honestly no longer see them as valuable. Probably sooner rather than later, solving complex optimization problems will become a forgotten art.
In its place, a reminder that I was once okay at coding.
Adrian Kajda retweeted
Android has always driven innovation in the industry through its unique flexibility and openness. At this important moment, we continue to be at the forefront of how developers distribute their apps and games on billions of devices globally.
Today we are announcing big changes:
– Leading the way in store choice: A registered app store program that makes installation of participating app stores easier with a streamlined flow
– A new business model for Google Play with lower prices for developers: The ability to use your own billing alongside Google Play Billing, and lower service fees.
These changes will make for a stronger Android ecosystem with more choice for users and even more successful developers who can further invest in higher quality experiences on mobile and beyond.
The above will be rolling out in phases, subject to local laws and processes.
We are also excited to announce that we’ve resolved all our disputes with Epic Games globally!
Details on all of this can be found at our blog post.
android-developers.googleblo…
Adrian Kajda retweeted
Anthropic bans their subscription from being used for API credits in ClawdBot, then forces a name change so it doesn’t sound like Claude, then Kimi K2.5 drops, open source, and is supposedly BETTER than Opus 4.5 anyways…
Couldn’t have set themselves up any worse than this
The name Clawd was only working in their favor - when I heard that, I assumed it was a harness only around Opus 4.5
The average Clawd user is going to default to an open source model now
Certainly one of the business decisions of all time
Adrian Kajda retweeted
The amount of crap I get for putting out a hobby project for free is quite something.
People treat this like a multi-million dollar business. Security researchers demanding a bounty.
Heck, I can barely buy a Mac Mini from the Sponsors.
It's supposed to inspire people. And I'm glad it does.
And yes, most non-techies should not install this.
It's not finished, I know about the sharp edges.
Heck, it's not even 3 months old.
And despite rumors otherwise, I sometimes sleep.
Adrian Kajda retweeted
Blaming the platform when your app can’t process a subscription is wild.
So many posts like this trash Android development without context.
👉 The Android app here was literally broken and didn’t even let users pay.
Community notes is the best 😆
Adrian Kajda retweeted
Sideloading is fundamental to Android, and it's not going anywhere. As we said in our blog, our new developer identity requirements are designed to protect users and developers from bad actors, not to limit choice. We want to make sure that if you download an app from a developer, regardless of where you get it, it's actually from them. That's it. goo.gle/3GUXYzT
Adrian Kajda retweeted
Mój pierwszy widok na nią i na Was. Dobrego wieczoru 💙
My first view of her and of you. Have a good evening.
@Axiom_Space @SpaceX @esa @MRiTGOVPL @POLSA_GOV_PL
#Earth #ISS #IGNIS #Ax4 #astro_slawosz
Adrian Kajda retweeted
Most apps submitted get rejected or approved in a day.
Apple has been holding Epic’s Fortnite submission for day 6 now -that uses web payments a nudge ordered Apple to allow in the Epic vs Apple lawsuit.
Apple clearly trying to block Epic for as long as they can. Playing dirty.
Adrian Kajda retweeted
Stripe 🤝iOS developers
Big news iOS app developers! You can now accept payments with @stripe outside of your app, with no iOS app store commissions.
The Stripe team 🧑🍳cooked up a quick guide walking you through how. Go live now!
Docs in the 🧵
Adrian Kajda retweeted
Nie jestem z @pisorgpl ani innej żadnej partii - moja narracja jest taka sama i takie wypowiedzi są po prostu szkodliwe. Te restrykcje nie powinny dotyczyć Polski jako jednego z największych “dostarczycieli” umysłów do rozwoju #AI na świecie, dodatkowo nie o same chipy tu chodzi.
Po wielu rozmowach i aktywności polskiego biznesu oraz branży tech większość amerykańskiego tech biznesu uważa te restrykcje za szkodliwe - kosztowało to nas/środowisko wiele pracy, więc prośba żeby się już najlepiej po stronie polityki na ten temat nie wypowiadać.
Z moich informacji również wynika, że albo w tym roku wrócimy do tier1 albo te restrykcje zostaną zniesione. Także niech już politycy nie ruszają tego najlepiej.
Adrian Kajda retweeted
Jednak ta mapka okazała się prawdą! A my w Polsce nie powinniśmy myśleć o jakiś globalnych zastosowaniach AI.
Na podstawie decyzji administracji prezydenta Bidena, Polska jest objęta restrykcjami w imporcie najnowszych chipów NVIDIA. Ostateczny dokument jest dostępny tutaj: public-inspection.federalreg…, choć pewnie lepiej spojrzeć na jego podsumowanie tutaj: whitehouse.gov/briefing-room…. Reguły te na pierwszy rzut oka wyglądają może niezbyt strasznie, bo limit 50 tyś kart jaki dostała Polska może brzmi jak dużo dla nas, a i jeszcze może zostać podniesiony do 100 tyś.
Jednakże jak ktoś wie co się dzieje w rozwoju AI i, że teraz forpoczta walczy o stworzenie superinteligencji. To wydźwięk tego dokumentu jest jednoznaczny - żółte kraje powinny zapomnieć o rozwoju własnej superinteligencji. A limity mają im wystarczyć tylko do zwykłych zastosowań AI bo "ten limit zapewnia, że amerykańska technologia jest dostępna do obsługi zagranicznych rządów, dostawców usług medycznych i innych lokalnych przedsiębiorstw". Czyli broń boże nie powinniśmy myśleć o jakiś globalnych AI. Żałuje, że ten wpis jest tak negatywny, ale ja osobiście nie rozumiem jak w tej całej konfiguracji geopolitycznej znaleźliśmy się w drugiej kategorii państw.
