Macilias (@Macilias) header
Macilias (@Macilias) profile picture

🇵🇱 NAFO

Germany
Joined December 2008
Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)Photo by Macilias (@Macilias)
Our first three ads for fak10, now in 🇬🇧 Really filmed, the real app, my own music Only the narrator's voice is AI. Here's Spot 3: Who is Ingrid? All three: fak10.app/en/blog/three-spot…
1
114
Macilias retweeted
This is wonderfully funny Thank you, Poles
🇵🇱🇷🇺🇩🇪🇮🇹🇫🇷🇬🇧🇺🇦🇷🇺 Podobno polska delegacja na rozmowy pokojowe z Rosją ma zażądać zwrotu Smoleńska, powołując się na pokój w Polanowie z 1634 r. Argumentacja prawna jest mocna. Pokój był wieczysty, a wieczność jeszcze nie upłynęła. Rosyjskie zastrzeżenia dotyczące późniejszych traktatów zostaną odrzucone jako wybiórcze podejście do historii. Skoro o granicach mają decydować dawne mapy, Polska również przyniesie swoją. Mamy ładną, z XVII w. Delegacja zapewni przy tym, że rozumie rosyjskie obawy o bezpieczeństwo. Dlatego po zwrocie Smoleńska Rosja otrzyma pisemne gwarancje, że Polska nie zamierza posuwać się dalej. Gwarancje będą dokładnie tyle warte, ile rosyjskie. Rosja otrzyma też zapewnienie, że przyłączenie Smoleńska nie jest wymierzone przeciwko niej. To wyłącznie odpowiedź na polskie obawy o bezpieczeństwo. Liczymy, że w Moskwie docenią znajome sformułowanie.
168
1,930
43
15,489
391,870
Macilias retweeted
Orędzie gubernatora RP.
79
685
47
4,424
461,214
🚨🇵🇱 Poland's constitutional crisis just took another rather EXTRAORDINARY TURN President Karol Nawrocki was reportedly ALMOST BLOCKED from landing in Warsaw after abruptly turning his plane around to confront the crisis at the Constitutional Tribunal. According to reports aired on Polish television, someone allegedly wanted his presidential aircraft denied access to the military section of Warsaw's airport and diverted to another city. The remarkable part: journalists claim the MILITARY REFUSED to follow orders from above, allowing the president to land. Nawrocki did make it back to Warsaw, where he confronted the escalating dispute over the country's highest constitutional court. Now the rather serious question is WHO allegedly tried to prevent Poland's own president from returning to his capital. And whether such an order was ever actually issued. Source: TV Republika / Writer: Oliver
362
4,881
138
22,122
487,504
Anna, die Micro-KI von fak10, erledigt den lästigen Papierkram. Sei dabei, wenn aus Neuland Zukunftsland wird. fak10.app/beta
2
3
322,362
Życie jest cudem. Dla każdego człowieka posiadającego sumienie. Naprawdę, nie musi być katolikiem
54
309
9
1,764
18,247
Czapki z głów dla @KinasRemek - bez niego nie było by Bielika i pierwszego polskiego modelu z reasoningiem trenowanego przez RL - rozwija biologiczne AI w świetnym polskim startupie (ingenix.ai) - po godzinach zbudował model decyzyjny lepszy i szybszy od oryginału (Jev) - z zerowym wsparciem, z własnych środków - doradza krakowskiemu SATIM-owi: buduje pipeline'y do wykrywania statków w portach i samolotów na lotniskach na zdjęciach radarowych z satelitów- Współautor publikacji Pathway o architekturze BDH - wielu młodych wyjadaczy AI go podziwia - do tego skromny, zawsze sympatyczny - bardzo zależy mu na polskim AI. w Ameryce byłby pewnie szefem dużego AI Labu Kwintesencja hasła „Polak potrafi” 🇵🇱 nitter.cf/KinasRemek/status/2111…
Replying to @carc_ai
Yes, it is my side project, paid from my private account, running training on paid @runpod (I ask them for help and sponsoring - they have not even responded with 'no'). Everything in this model is my experience from my all previous AI projects.
38
301
8
2,600
59,204
Polak potrafi. Także i cyfrowy.
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.
2
107
It’s 2026, and people are still criticizing @elonmusk for pursuing ambitious projects that supposedly make no sense. This time, the debate is about flying cars - whether anyone actually needs them and all the practical problems they might create. So let me explain this once again. Whether it’s a mission to Mars, a next-generation vehicle, a flying car, or any other seemingly unrealistic ambition, the ultimate value of these projects goes far beyond the products themselves. Ambitious projects are catalysts for extraordinary engineering breakthroughs. To make them possible, we need innovation on an unprecedented scale across materials science, chemistry, physics, mathematics, computer science, artificial intelligence, robotics, energy storage, thermodynamics, aerodynamics, advanced manufacturing, control systems, telecommunications, and countless other disciplines. These challenges force us to solve problems we would otherwise never attempt to solve. They push the boundaries of what is technically possible, generate new technologies, create entirely new industries, and accelerate scientific discovery. And crucially, the resulting innovations rarely remain confined to their original purpose. They spill over into other sectors, improve productivity, reduce costs, and create economic value far beyond the initial investment. The point is not whether humanity desperately needs a flying car. The point is what humanity must invent to make a flying car possible. Of course, you can disagree with Elon Musk’s worldview, his political opinions, his decisions, his communication style, or many other aspects of his personality and business practices. That’s perfectly reasonable. Admiring technological ambition does not require agreeing with the person behind it. But we should be able to separate our opinions about individuals from our appreciation of the engineering challenges they are willing to take on. We should also stop evaluating ambitious engineering projects exclusively through the lens of their immediate commercial utility. Not every ambitious project will succeed, but even failures can produce knowledge and technologies that become the foundations of future breakthroughs. Sometimes the greatest value of building something extraordinary lies in everything we discover along the way. And one thing is certain: what an extraordinary time to be alive, to witness breakthroughs that previous generations could only dream of. We should appreciate that more often.
46
40
6
581
30,463
You’re so worried about AI music but all you do today is sample old songs!
709
2,151
159
17,026
408,585
Macilias retweeted
Ej, @pozar_w, udało się wyciszyć?
68
366
3
2,779
28,813
I know it, and it ain’t yet another LLM 🤗
America has ChatGPT. America has Claude. America has Gemini. America has Grok. America has Llama. America has Muse. America has Dots. China has DeepSeek. China has Qwen. China has Kimi. China has GLM. China has Hunyuan. What does Europe have?? 🇪🇺
1
140
Next episode introducing Anna, This time focus on Relation Typs.
Devlog 16: How Anna understands relations. Say "Mila is older than Jonas" and she also knows Jonas is younger, stored once. She infers chains, asks when something doesn't fit, and drops inferences once their basis is gone. Anna's voice: AI · fak10.app
1
118
Macilias retweeted
Nada como ver una tendencia con esta joya de video 👌
134
3,887
217
27,886
994,786
Macilias retweeted
Replying to @BaT_MaN_23
Done.
4,665
26,453
4,807
238,200
13,788,232
Big milestone this weekend. The navigation within Anna’s Pensive (Ontology) is finally sound 🤗
An inspiration for fak10: Dumbledore's Pensieve. Memories rest side by side, and when you dive in, you see how it all connects. For Anna that's her knowledge store, her Pensieve, your ontology. Tap a name and wander through everything you've told her. Anna's voice: AI
130
How does remorse feel to you? I'm teaching Anna what feelings are: 16 feelings, two clicks each, anonymous, ~3 min. At the end you get your own feelings map. Save it: when the survey closes, I'll publish everyone's average to compare. fak10.app/en/feelings.html — Maciek
1
1
43
Ich implementiere grade die Gefühl Logik für die App fak10, und bräuchte da ein paar Freiwillige für eine kurze anonyme Umfrage. Dauert echt nicht lange und zum Schluss poste ich auch die Mittelwerte, da könnt ihr selber gucken wie ihr da steht 🤗 fak10.app/gefuehle.html
2
115