Generating all software on the fly is obviously not the goal here - it would be terribly inefficient and inconsistent. But if we could adapt interfaces to our needs or build new features on the fly, share it with colleagues for feedback, and merge broadly value-generating changes to the core, imagine how user-centered software could naturally and organically become!

AI might not just kill coding...it could swallow software itself. With @Claude Sonnet 4.5 you describe what you want, and software appears in real time. That means an endless flood of apps, systems, and tools... until “apps” stop meaning anything at all. And it all happens inside what looks like a simple chatbot, where conversation replaces code. At first this just means an explosion of apps, systems, and tools. Anyone can build anything instantly. But when software is this abundant, the meaning of “apps” begins to disappear. Execution stops being the bottleneck. The scarce resource becomes judgment: deciding what is worth building, and why. Over time, that shifts the whole system. Education no longer trains us to memorize but to think critically about goals, ethics, and design. Work stops revolving around pushing buttons on screens, because the screens themselves fade. AI assistants, tied into robotics and infrastructure, carry out most tasks automatically. The computer as we know it becomes invisible, woven into life itself. We’ll still have challenges, decisions, and tradeoffs. But the nature of “work” may change from doing tasks to governing abundance. So maybe the real endgame isn’t more apps, but fewer. Not more screens, but less. A society where the computer disappears, and the questions left are: what do we build, what do we ignore, and what do we protect? What do you think?

With tools like Copilot and Claude, “vibe coding” makes development faster than ever, but speed comes with risks. Over-reliance, hidden flaws, and AI “hallucinations” can cost more than they save. Focus on how to harness the power of vibe coding while avoiding its pitfalls will be crucial in the near future https://lnkd.in/dfiMK55D

AI programming’s worrying stat: 70% of developers trust the output. We need an intervention. New research from DORA confirms AI is ubiquitous. 95% of developers rely on it, and 80% report productivity gains. But there is a massive and dangerous disconnect. That trust is simply not supported by the facts. External research shows AI systems routinely make unsupported or fabricated claims, with failure rates ranging from 30% to 95%. When AIs are overeager and rush ahead with little insight, relying on them without scrutiny is a recipe for disaster. AI can and often will change the entire implementation after a small change, putting us in grave danger of losing the deterministic foundations of programming. We need to treat AI as an imperfect collaborator, not an infallible oracle. The core principles of software engineering are now MORE important, not less: 1. Work in small steps: actively constrain the AI. 2. Verify everything: after every small change, verify that the system is still working as expected. Successful AI adoption takes discipline. If we abdicate our responsibility to accurately specify and verify what we’re building, it is "unlikely to end well".

I think one of the most difficult and tiresome process now is testing AI-generated code given how much code AI generally spits out. This then leads to the mentality of just "trusting" the AI to code well.

If AI can generate an entire book every hour, who will be able to edit or even read it? There are speed limits on most roads, despite the capabilities of the car. Although I don't know what subjective measurements Dora used this time, I do agree with the sentiment. I would go a bit further, though. - The AI circular prompting hell is a massive waste of resources and the human mind - Companies have to be vigilant about quality - Generating more doesn't mean better - Generating more will create a bottleneck and strain

In using AI to generate code "There are some big productivity gains for developers who use the tools wisely. But for those who yolo their way through... it's going to be ugly" Eric Minik

Why do many fail using ai for coding? they expect perfect code from a single prompt and spend nights debugging the 'ultimate nothing.' reality is ai suggests, but doesn't replace deep understanding. just giving a prompt and hoping aI will do all the work won't cut it. you need to guide it step-by-step, review results, and know where to put your hard work, like system design and in detailed description, asking questions. ai helps speed up parts of coding, but tough problems still need your brain. quick fixes don't build strong software. mastering fundamentals and system thinking and most imporant ask question to ai utill you get clearity. don’t rush to get fast results. take time to craft good prompts, understand ai output, and combine ai help with your skills. this saves hundreds of hours of debugging pain later. remember ai is your assistant, not a magic wand. use it smartly, do the tough work yourself, and you'll not only write better code but grow as a developer.

💡 “If AI can write code, does it still make sense to learn coding?” It’s a fair question. AI tools today can generate functions, fix bugs, and even suggest full solutions. But here’s the thing: AI can write code — it cannot replace the mindset of a coder. When you sit down to code, you’re not just producing lines of logic. You’re training yourself to think systematically, to approach problems creatively, and to keep going when the first (or tenth) attempt doesn’t work. Those habits are what shape great developers — and they go far beyond syntax. At Codepresso, we see this every day. Coding practice builds the resilience and clarity of thought that no AI can substitute. AI is an accelerator, yes — but coding is the foundation. Without understanding why something works, you can’t truly innovate or adapt when things go wrong. So my answer is always the same: 👉 Keep learning, keep coding. AI is here to partner with you — not to replace you. Hyunjin Hwang SEOKMIN JANG Jihoon Kim Saebyuk Lee Ellie Nguyen Kristel Weinreich António Ferreira SkyinQ Soyeh Kim Irene Mäkelä-Brunnekreef Jasurbek Ayupov Jake Kim