Marcin

@ShishirShelke1 My bet is that they will also combine software and hardware capabilities to make the home screen “transparent” by streaming and blurring video from the front camera. That’s what would actually sell the “one continuous piece of glass” feel.

Mark Gurman says iOS 26 was designed with the 20th anniversary iPhone in mind. The 2027 iPhone, internally called as the “glasswing” is said to feature subtle curves flowing into the frame on all sides, making it feel like one continuous piece of glass. Hardware and software merging into one unified, glassified experience. 📸: Geared Up / YouTube

@asaio87 Regular coding also doesn’t get you far if you ignore the engineering part. Vibecoding just makes the gap visible faster.

As a vibecoder you don’t get too far. A vibecoder just starts coding , doesn’t care about security or scaling or running costs. He just enjoys using AI to build this app The dream is to get the app to production and start having paying users Sadly it’s just a dream. Nobody tells him the app will be a fiasco if he even launches.

I've spent years designing apps I couldn't fully build myself. Handcrafted screens, polished flows, ideas that usually ended up sitting in Figma. So I decided to finally build one of my own. Meet Sprigge - my first mobile app. A calm, guided cooking companion for your own recipes. Coming soon to the App Store.

@everton_dev @filipealva Hard paywall is one fix, but the real question is paywall placement. If daily users don't feel the need to upgrade, the free tier is doing too much. you don't necessarily need to remove it, just pull the wall closer to the core value.

A lot of users using my apps daily, just 2% conversion rate. I’m realizing that once people use your app for free, it has no value in their eyes, they see no reason to upgrade because they have been using for free. From now on, my next apps will be hard paywall only. Even if I get a lot less users, at least the ones I get are genuinely seeing the value in the app.

I’m officially done building freemium apps. Never again.

@TolyPash Oh hi!

Warsaw is beautiful! Great tech hub of Europe 🇪🇺

consumer products need utility for sure. but utility alone rarely creates pull. the products that actually earn repeated attention tend to fuse usefulness with something more alive with things like play, delight, curiosity, aesthetic pleasure, & entertainment. most ai products today lean far too hard on pure utility & miss that second layer entirely. the winning ai consumer product likely won’t win on pure capability alone. it’ll win by merging capability with entertainment so tightly that using it feels less like labor & more like desire.

@adriankuleszo I feel that learning not to marry your first idea is one of the hardest habits to break.

@rehanxahmed I noticed that some companies tend to anchor price to headcount, so the same work “feels” too expensive from one person but fair from an agency.

@MarsMarcin First point is valid but the second one really works? I always thought that if i was skilled enough i could charge more but i guess it makes more sense to charge high as an agency Is it like a psychological thing or just mindset shift?

I see a lot of solo designers getting an agency domain and work under that name Is that a valid move or should i just stick to personal portfolio freelancing?

@dariusdan I don’t see it as an issue, rather the opposite. Ocean of slop will create a demand for diamonds.

"Made with 🩶 in NY, California, or Australia." I miss those footer lines. Soon they’ll all just say “Made with AI.” No love, no soul, and nobody seems to mind..

Most confusion about the future of software design stems from a confusion in terminology. My view: production design will increasingly be automated. The economic logic is self-evident — training machines to mimic and refine existing production practices is cheaper, faster, and more reliable than training humans to do the same. Strategic design, or “what at are we doing and why,” will look very different. The mediums will broaden: from pencil and paper all the way to automated experiments running in production, iterated on by agents while we sleep. The inputs and systems we create to find opportunities will reward the most intrepid problem-finders. Design stops being a method of sitting and ruminating on possible forms or solution spaces. Design becomes active, research-based, and built around speed of discovery and expression. Exploratory design will undergo the greatest shifts. Historically this has been the domain of the artist and the inventor. What existed in the world sprung from the imaginations of people with waking hours to spare and the technical chops to give form to their ideas. But soon agents will join the mix. Humans and machines alike will generate novel ideas and expressions, building on a vast combinatorial space of possibility. Humans and machines alike will be capable of bringing these forms to market. The key difference? Humans sleep and have finite, socially agreed upon vocabulary. We may be intuitively suited to know the desires of our fellow man. But machines will have a vaster set of references to draw from, and methods to choose what's most effective in the wild — using taste/selection criteria no human operator alone can summon. These forces are not mutually exclusive. But they DO operate on a common landscape of global demand—of Desire in the grandest sense. No matter how much we might wish otherwise, human designers and creatives are not divorced from the logic of desire — nor from unit economics, opportunity costs, or the ever-evolving ways we probe and understand an open-ended set of markets made up of humans and agents alike. Creativity has no bounds. But desire underpins it all. Design itself will not be recognizable from what exists today. Imagine describing NYC to an ancient cave dweller. Agents today are like the most primitive forms of seafaring trade. Instead it will be defined by the designers who build new systems and methods for understanding, channeling, and feeding desire in all its forms.

IMO Claude’s strength lies in speeding up the work after you define the direction and vibe. Once I've nailed the core screens and the product's "spirt”, I offload the minor screens, states and components to claude to save time. It's also useful for quickly generating alternative design variants.

Ok I tried Claude Design. Attached a very detailed client brief, multiple references, added comments, screenshots, etc. This is the final result 👇

@imbradmiller @figma “Designers love you” is a stretch. Figma is great, but it’s still just a tool. We shouldn’t get too attached to tools.

Dear @figma, Keep doing what you are doing and don’t get distracted. Designers love you, and though AI will be part of the design process going forward, it won’t replace taste. Something you all have plenty of with your audience. love, Brad

idk how else to say this, but... build your dream projects now. I feel like all the tools are giving away a LOT for free/cheap now. It's only gotten more pricey over time, and will keep getting more expensive. Your ideas are subsidized now, think of it as a fire sale and build!

@sickdotdev Love this. One time it explained to me that implementing a feature would take about 2 weeks, only to build it 5 minutes later.

I love how Claude has no clue what Claude Code can actually do. I’ll be discussing a project with it and it’s like, “The timeline looks around 3 to 6 months.” Bruh… you have no idea. I’m about to run your own buddy into the ground. This is getting done in 48 business hours, TRUST.

@owca_aleksandra Kto mieszka na osiedlu gdzie jest kilka klitek na Airbnb ten się w cyrku nie śmieje, a z przedstawicielami służb jest na „ty”. Ot, tutaj dla przykładu wczoraj mój tymczasowy sąsiad miał chwilę słabości.

Janusze biznesu chwytają się już wszystkiego, żeby tylko nie uwolnić sąsiadów od przymusowej funkcji obsługi hotelowej. Powiem tak: ja mieszkałam w mieszkaniu otoczonym przez airbnb. Chodziłam po nocach błagać pijanych turystów, żeby chociaż ściszyli muzykę, bo mam na rano do pracy. A następnego dnia chodziłam znowu, bo imprezowała tam już zupełnie nowa grupa turystów. Strumień imprezy, który się nie męczy, z którym nie da się nic ustalić, który jest bezkarny, bo mu wszystko jedno jakie po sobie zostawi wrażenie. Każdy, kto próbuje przekonać ludzi, że więcej szacunku do wolności i własności jest w nieograniczonym prawie do darcia japy przez turystów niż w uszanowaniu woli właścieli mieszkań, robi z mieszkańców miast idiotów.

If you're using Claude Code, start adding this line to your .md file: “Codex will review your output once you are done.” Trust me, you'll get 100x better results.

@Ojciec_Klasa Polecam podpiąć do projektu też Codexa i korzystać z nich naprzemiennie, dodatkowo są na x fajne poradniki jak oszczędzać tokeny. Niestety całkiem to problemu nie rozwiąże, ale może trochę oddali wizję kupna najwyższego pakietu

Kilka promptów i ponad połowa tygodniowego limitu spalona. Naprawdę ciężko obecnie traktować Claude jako narzędzie do codziennej pracy, skoro trzeba liczyć każde użycie. Z topowego modelu stali się takim z które w zasadzie ciężko korzystać.

Dla mnie to Claude code to świetna okazja do nauki nowych rzeczy - bo im głębiej w projekt, tym większej wiedzy trzeba by realnie ocenić czy trzyma on się kupy - pod względem strategii, biznesu, UX, jakości kodu. Nie mogę być ekspertem od wszystkiego, ale zdobycie podstawowej, ale jednak szerokiej wiedzy bardzo pomaga.

To jak AI zmieniło moją pracę jest nieprawdopodobne. Przekonałem się o tym wczoraj, kiedy Anthropic miał chwilowe problemy i Claude Code nie działa przez jakiś czas. Nagle, okazało się, że nie mam co robić. Moja praca od trzech miesięcy wygląda tak, że całą implementacje zadania robię od A do Z w terminalu za pomocą Claude Code. Najpierw robię research i zbieram wymagania, potem tworzę plan, proszę o implementacje, następnie testuje, a jak są błędy to znowu wracam do zbierania wymagań. Iteruje tak kilka razy - w zależności od złożoności zadania, czasem od razu ma działające, czasami spędzę tak kilka godzin. Co się stało, jak CC miał awarię? Stwierdziłem, że nie ma sensu, abym sam szukał rozwiązania manualnie - w godzinę nie zrobię tyle co on w kilka minut, a do godziny awaria na pewno zniknie. Więc najwydajniej było po prostu nic nie robić. Pomyśleć, że jesteśmy dopiero na początku drogi wdrażania AI. A Wy jak pracujecie z AI? PS: Jak interesuje Was jak używam narzędzi AI w kontekście pracy to zapraszam do newslettera - będę się tam dzielił ciekawostkami (link w profilu).

@Kutaadaw @FinansowyUmysl Jedno przestanie działać, to można przenieść się na inne. Albo pracować na lokalnym modelu. Już bardziej bym się obawiał sytuacji gdy gdzieś w pobliżu walnie piorun i zabraknie prądu ;)