Rob

@pcshipp please charge your phone

I have no words for this, wtf this calculator makes $400K in revenue No API cost, no server cost, no infra cost

@levelsio @TermiusHQ /remote-control and send the screenshots from the Claude App!

So yes the #1 problem I have with my VPS Claude Code set up is copy pasting screenshots into @TermiusHQ, if someone can fix that I'd be very happy Termius themselves should just fix this: - detect paste of image - immediately upload it via SFTP to /tmp on server - show a progress bar - paste the /tmp/filename.png into the chat Please Termius make this!

@TomOMalleyX @levelsio /remote-control and send it through Claude app

@levelsio how do you give it screenshots over ssh

I did (SSH)

@primodelbarrio Es que es simplemente mentira que haya 4 millones de casas vacías en las grandes ciudades... Partes de una premisa falta y te pones a hacer malabares macho Más de la mitad está en municipios de menos 50k habitantes

Que haya casi 4 millones de viviendas vacías (muchas de ellas en grandes ciudades) es el síntoma de un modelo de país que hay que corregir atacándolo desde diferentes flancos pero con la mirada puesta ahí: - Turistificación: Miles de casas se han sustraído del mercado residencial para el turismo, reduciendo la oferta y disparando el precio de los alquileres. - Derecho al arraigo: En los pueblos, el vacio es fruto de un modelo que obliga al éxodo por falta de servicios. El reto es que nadie TENGA que irse de su casa. Proponer 'construir más' como única via es repetir la receta de la burbuja de 2008. Menos ladrillo y más gestión de lo que ya tenemos.

@gabriberton just do /simplify at the end of each coding session

Vibe coding creates lots of dead code. Run this often. You're welcome --- Delete all dead code. Use ruff and vulture ---

@esjavieraleman Paraíso fiscal dice el menda por tener un IRPF de 45% máximo en vez de 50%

Porque Madrid es un paraíso fiscal y usarlo de ejemplo es de mongolo

@zuess05 SaaS are not that easy to replicate, especially prod ready apps with customers. Most apps costing 29$/month really do something you can't simply one shot, you'll find it out after months of working on it

Genuine question. If Claude can clone your entire SaaS in 45 minutes, why would anyone pay you $29/month for it? What is your actual moat?

@BoringBiz_ no single actual tech person uploaded those videos..

Tech workers realizing that they could have kept high pay, job stability and remote work if they just stopped making cringe “Day in my Life” videos on TikTok

@Ranicket @sporadica rest of the world probably won't

@sporadica False. I think India would be very very happy to be part of the USA. Russians, on the other hand…

i shit you not, the most ideal solution to this is probably keeping India and China as is, then make literally all the rest of the world apart of the United States of America

@VicVijayakumar G - you for Spain

There’s one obvious choice and no one in their right mind would pick anything else.

@gymbroinvestor A tope tío, esos números son de envidiar en este país. Si buscas otra empresa para afiliación, en @allinvestview buscan gente como tú!

Resumen ingresos y gastos marzo 📊 INGRESOS 📂 Excel → 140€ 🚀 Afiliación → 3.316€ 👨🏻‍💻 Proyectos → 2.250€ 🫂 Promoción → 3.001€ 🎯 Monetización → 91€ 📈 TOTAL → 8.798€ GASTOS 🛍️ Ropa → 198€ 🎹 Piano → 63€ ⛷️ Viajes → 411€ 🥩 Comida → 372€ 🚙 Gasolina → 141€ 🏋🏻‍♂️ Gimnasio → 54€ 👨🏻‍💻 Proyectos → 1.658€ ✍🏼 Autónomo → 88€ 📉 TOTAL → 2.985€ INVERSIÓN 🟡 Oro → 535€ 🟠 Bitcoin → 450€ 🇺🇸 S&P500 → 1.500€ 🏡 Inmobiliario → 500€ 🤖 Roboadvisor → 250€ 🌍 Emerging Markets → 250€ 📈 TOTAL → 3.485€

@manuelabergerot Ni un político con sentido común hay en este país. En esas casas ya vive gente, inútiles. Por qué no proponéis construir nuevas viviendas en vez de robar al actual propietario?

Ayuntamientos, comunidades autónomas y Gobierno están tardando en adquirir estas viviendas para convertirlas en alquileres asequibles. Mientras tanto, que los mega caseros cumplan la ley. Si tu contrato de alquiler vence antes del 31 de diciembre de 2027, manda el burofax.

@Bfaviero @TheStalwart or say something like we've been discussing this for 2 hours, outcome is not gonna change, send it now. (he makes up the time though)

@TheStalwart I once asked it to further refine an email and it said "just send the damn email"

I asked Claude Code what an easy way to work on a project is while I was on vacation without a laptop and it said to just enjoy the vacation

@aakashgupta how it's different from asking in another session? does it respond as well with the full context of the current session? If so that's huge

The biggest bottleneck in AI coding tools has always been session architecture, not model intelligence. Every agentic coding tool today forces the same workflow: give a task, wait, watch, twiddle thumbs. You either sit idle for 30 seconds while the agent works, or you interrupt it and blow up the context window. /btw fixes this by treating a Claude Code session like an OS instead of a chat thread. The main process keeps running. You open a side channel. Ask your question. Get your answer. Main task never flinches. This matters because the average agentic coding session runs 15-40 tool calls per task. Every interruption risks derailing that chain. Developers learned to batch their questions and wait, which is exactly the wrong workflow when you’re paying for a tool that’s supposed to save time. The comparison to make is threading. Before multithreading, CPUs ran one process at a time. The hardware was capable. The architecture was the constraint. Same dynamic here. Claude’s model can handle parallel reasoning. The session layer was forcing everything into a queue. Watch how this compounds. Side conversations mean developers stay in flow state during long refactors. Flow state means better oversight of what the agent is producing. Better oversight means fewer reverts. Fewer reverts means the whole session completes faster. Anthropic keeps shipping features that treat developer cognition as the scarce resource. That’s the right optimization target.

@ChrisLaubAI Why are we assuming IA is 100% autonomous? It's definitely not there yet, but add one human in the loop and you have extraordinary speed and still good quality code

BREAKING: Alibaba tested 18 AI coding agents on 100 real codebases, spanning 233 days each. they failed spectacularly. turns out passing tests once is easy. maintaining code for 8 months without breaking everything is where AI completely collapses. SWE-CI is the first benchmark that measures long-term code maintenance instead of one-shot bug fixes. each task tracks 71 consecutive commits of real evolution. 75% of models break previously working code during maintenance. only Claude Opus 4.5 and 4.6 stay above 50% zero-regression rate. every other model accumulates technical debt that compounds with every single iteration. here's the brutal part: - HumanEval and SWE-bench measure "does it work right now" - SWE-CI measures "does it still work after 8 months of changes" agents optimized for snapshot testing write brittle code that passes tests today but becomes completely unmaintainable tomorrow. they built EvoScore to weight later iterations heavier than early ones. agents that sacrifice code quality for quick wins get punished when the consequences compound. the AI coding narrative just got more honest. most models can write code. almost none can maintain it.

@gabrielrufian No puedes ofrecer soluciones cuando la premisa es falsa. Las familias compiten con boomers, no con fondos. ¿Cuál es tu solución real?

Ni reformar más. Ni construir más. Ni subvencionar más. INTERVENCIÓN del mercado. El problema se llama: ESPECULACIÓN. Y no se soluciona con parches, se soluciona gravando y PROHIBIENDO la compra especulativa de vivienda. Que las familias dejen de competir con millonarios o con fondos millonarios. Quien quiera hacerse rico que se compre un cupón de la once pero no una vivienda. Una FAMILIA, una CASA.

@ClaudiusMaxx @itsandrewgao I think the model already knows what it is most of the time

better move: dump component.gallery into a markdown file and drop it in your agent's context. now every prompt gets the right vocabulary without you memorizing anything. the model knows what a "skeleton loader with shimmer" looks like. you just have to remind it those words exist.

you can instantly 10x your vibecoded frontends by just learning what different ui components are called ofc opus is creating generic slop, the only words you know are menu and button.

@levelsio For me the most important issues are regressions, it may add a new feature but change something that silently broke something else on the app. You definitely need to check what he's updating because it will make an stupid code change sooner or later

This week I decided to just permanently switch to running Claude Code on the server mostly on bypass permissions mode: c() { IS_SANDBOX=1 claude --dangerously-skip-permissions "$@"; } And for the first time in my life I think I've actually managed to outrun my todo list What happened is I simply blasted through my to do list of features I had to build and bugs I had to fix I've never shipped so fast and Claude Code almost made no mistakes, and when it did it they were tiny that weren't fatal (important because I'm mostly working on the server in production now) Before I was always known to ship fast (also because I always work alone) but while I shipped new things would always build up on my features/bug board (my users can submit them there) But this is the first week where I've been fast enough to outrun them The board is actually empty! As other people have written on here the real bottleneck is becoming myself and my creativity, not how fast I can ship. Because I think I ship faster now than I can come up with new ideas, or maybe my brain will adjust to this new speed (probably) Also I feel another limit is becoming my own mental context window, as in how many things, features, bugs, projects, I can keep in my mind in parallel while building on all of them. It's a lot and I haven't reached that limit yet but I feel I might be close I also noticed that you start going really fast the more you let it just go loose, before I was slow because I didn't trust it and I was scared it would destroy my code, now I just let it go. As @karpathy wrote, things feel like they've changed a lot around December last year when models became good enough to really code with and I feel the same When I see other friends code with Claude Code I often notice they're slow because they still check everything, which is good of course, but I feel the better way would be to create some tests and just let it run freely and see if it can pass those For me the tests are mostly just me checking out if the new feature on the site works or not, and in 99% cases it just does, and then I ask it to improve it further Because I run Claude Code on the server in production, I don't have to wait for deployment anymore (although that took only 3 seconds anyway before, that still adds up), now it's wait for it to be done coding, I refresh the site and I test it, that feedback loop is how I work and it's made me WAY faster Anyway here's what I did this week and the majority of these things were requested by people on the bug board, I'd say this is about 10x my normal output: 📸 Photo AI - Built new image viewer and mobile image viewer - Added batch remix, multi-photo import, filtering by model in gallery - Security overhaul: phased out insecure ?hash= login, migrated to session tokens - Fixed Google login loop, multi-model selection, talking scripts - Added custom audio upload for talking videos - Created dynamic model selector from server endpoint 🏡 Interior AI - Revived [ Add furniture ] feature (started 6 months ago, image models now good enough) - Added custom style upload for redesigns - Built own Gaussian Splat viewer for 3D - Made /remove_bg endpoint for furniture backgrounds - Migrated 3D walkthrough to new World Labs API - Added .skp file support, paint color masking, empty room button 🎒 Nomads - Launched weekly AI-generated newsletter from chat - Built profile edit modal, moved profile editing from /settings to profile page - Added TikTok/YouTube links, status bar, server-side API tracking - Added hundreds of new profile tags and traits - Fixed timezone filters, broken links, user avatars 🗺️ Hoodmaps - Revived write mode (before was only read for last few years because db was rekt) - Built heatmap mode using sentiment-scored tags (50K+ tags) - Fixed root cause: tags not entering DB due to wrong PRAGMA (should be WAL) - Added good/bad area detection with admin grid controls - Set up Claude Code Telegram bot for live changes - Enabled CF cache, fixed health check, fixed Brussels 📕 MAKE book - Built auto ePub/PDF generator cron worker - Added dynamic generation with personal customer watermarks - Added image compression for file size 💾 Pieter .com - Added Wikipedia text-only reader for Kindle - Exploring Windows 3.11 emulator using v86 (to replace Em-DOSBox) - Added product recommendations on homepage - Installed Wall Street Raider (1986) 👩‍💻 Remote OK - Installed Chatbase AI customer support bot - Added "report not remote" link on job posts 🏨 Hotelist (3 todos) - Fixed hotel URLs and city range bugs - Added iron amenity

@sapeadorZ @7uanF la gente básicamente no sabe ni buscar en Google, pero ahora se va a poner a investigar temas que ni le interesan porque IA

@7uanF Desde hace más de 10 años que tienen el conocimiento del mundo en el celular, que las universidades son gratuitas o becadas y mil unas otras comodidades y... aún así, la gente prefiere hacer su vida simple. La condición humana,

Miren, les voy a hablar clarísimo: es tan fácil como preguntarle a la ia cómo hacerlo, sobre arquitectura de software e incluso sobre escalabilidad y seguridad ya paren de repetir esto como para asustar a la gente, este tren no lo detiene nadie les guste o no les guste

@Cynthia_67771 it is!

@rammanq What a lovely view👌

POV: you work in a start up