As I’m reaching the end of my robotics and AI engineering studies and starting to work on my thesis, I have two options, on the topic. My initial subject was supposed to be about my AI learning platform, which I am building currently. But now, as I’m diving deeper to learning about the folder structures and seeing AI in completely different daylight, which I haven’t seen anywhere else I feel like I should do it about that instead. What do you guys think?

Superwhisper Ultra runs a local speech model on my machine. No server, no upload. Nothing leaves the laptop. When I hit the hotkey and stay silent, the default transcription comes back as: "Thank you." Every time. Dead air in, gratitude out. 😆 It's my favourite bug, and I don't even know if it's a bug. Whisper-family models were trained on mountains of YouTube audio, and a huge portion of YouTube audio ends with "thanks for watching" or "thank you." So when the model hears silence, its best guess is gratitude. My computer thanks me for nothing, several times a day, and that is a mood. Now the serious part. A lot of people have shifted their dictation stack to cloud tools. Whisperflow, the cloud tiers of other apps, the rest. The pitch is speed and accuracy. The cost is your voice. Every time you dictate into a cloud model, your audio, a business conversation, a client name, a half-formed product idea, gets shipped to a server you don't control. Transcribed by a model you don't host. Logged somewhere you can't audit. This isn't a conspiracy. It's a trade. For casual writing, the trade is usually fine. For everything else, it's worth thinking about. And here's the tell. Tools don't usually get worse on purpose unless the maker wants you off them. Whisperflow has degraded for a lot of people lately. The cloud path keeps getting flakier. That's not random. Would you as a dev not fix something that users are saying has degraded... If your dictation tool is quietly worse than it was six months ago, ask what the why is... I moved to Superwhisper on the local Ultra model. Everything stays on my machine. Client names stay on my machine. yes, it's not quite as good, but yes, I know exactly where it's run and what data is being passed around. Code red ideas I wouldn't post about on the internet or talk about in public stay on my machine. And when I'm silent, it thanks me. Your voice is data. Keep it on the box you own // A<3

Many of us in the AI space are still selling the wrong thing. We package custom models, prompt engineering, agent setups, or hourly consulting. The conversation stays locked on tools, hours, and technical specs. By 2026 that model has clearly hit its limit. I’ve lived the shift firsthand while grinding open-source stacks for nonprofit work—Cloudflare tiers, Alibaba credits, Jake’s method, ZeroClaw experiments, and finally landing on Hermes Agent + Cognee. The tech itself became fast and cheap. Specialized work turned into a commodity almost overnight. Clients compare on price and speed, and the race-to-the-bottom kicks in. That’s why I flipped the script: stop selling the service and start selling the measurable value. Lead with real outcomes—less manual grunt work, tighter decision loops, lower operational costs, and repeatable processes that actually move the needle. Structure engagements around shared success metrics instead of hours or fixed deliverables. Price based on impact: performance bonuses, value-share pieces, or retainers tied to sustained results. When the system improves, everyone wins. Jake’s approach is a perfect example of this in action. He’s giving away his Interpretable Context Methodology for free, openly sharing the full system so anyone can adopt it. That single move creates exponential value across the entire community—more capable setups, clearer thinking, and faster progress for all of us without gatekeeping or hourly billing. Keep your own work minimal and transparent. Dig into the client’s real friction points. Co-define success in concrete terms. Build lightweight, maintainable architectures instead of black boxes. Show clear before-and-after results. The outcome is powerful: the system handles the repetitive load reliably, freeing up attention for the judgment and creative work only you (or your client) can do. You shift from being another vendor to becoming the guide who cuts through the noise. I’ve seen this mindset pay off in my own volunteer projects and early client conversations. It just feels cleaner and more aligned.

i keep switching between sites and still feel like i'm missing things. by the time i find something useful, it's already old news. looking for one solid source (or a tight stack) that covers : models, tools, updates. practical stuff i can apply. not just noise. tbh this community has been one of the better ones so far. but curious what else you're reading. website, newsletter, youtube, discord, anything. what are you using day to day?