... Claude before I even search the web. I use them to brainstorm ideas, review code, rewrite emails, summarize documents, or think through problems. They've become part of my workflow rather than just another website. I don't remember consciously changing my habits. It just happened over time. If you'd asked me a year ago, I would've said I'd never share drafts, snippets of code, or work-related notes with an AI. Have your boundaries changed ...

... recent data leaks and state surveillance laws, more people seem to prefer 100% local, on-device compute over cloud-based LLM architectures even if it means working with slightly smaller models. As users or makers, are you consciously choosing on-device alternatives

... surface the old decision if you ask for it directly. But in normal recall, the system leads with what is currently true. This matters most for: Decisions that evolve over weeks or months Preferences you have updated but not consciously logged Plans that changed after new information arrived The old memory also stays linked to the new one, so you can always trace why you changed your mind, not just what you decided. How are you currently handling stale context ... ... your notes or memory tools? Comment from Mustafa Arian(@thenameisarian): you already lead with the right instinct: keeping the old memory linked rather than overwriting. the follow-up i'd think about is when to surface the conflict visually vs

... useful. About 18 months ago I realized voice is the most underused health signal we carry. The same math I used on seismic waves applies directly to what your voice reveals stress, fatigue, energy, breathing all encoded before you consciously feel them. So I built HealthOS: 30 seconds of unscripted speech, 8 biomarker reads, fully on-device. No wearable. Your audio never leaves your phone. Won 1st prize at Health Wildcatters' 2026 TXHCC Hackathon. Pre-launch now with a TestFlight ...

... worth a look if your team is re-evaluating right now. Happy to share more if anyone's curious. Comment from fmerian(@fmerian): curious what are your preferred models? @MiniMax ? @Qwen ? else? Comment from fmerian(@fmerian): @bullterriertime  what alternative to @Github Copilot would you recommend? Comment from Debbie Shapiro(@debbie_shapiro): I'm not using Copilot, but invoices are coming to Claude starting next week as well. I built a tool (I launched here a week ... ... budget. It's challenging for sure. I've moved some agents to cheaper or local models, moved as many deterministic things as possible to scripts, etc. Comment from fmerian(@fmerian): "The harder question is whether to stay in the VS

... taught me something I couldn't un-know: our voice produces an enormous amount of signal every time we speak cognitive load, vocal strain, fatigue, stress, mood, and countless other markers all acoustically measurable, often days before you'd consciously name them. The problem is nobody built something ordinary people could actually use daily. So I did.HealthOS is a 2 5 second voice check-in. You just talk no script, no performance. It extracts signals from your voice: energy, fatigue ...

... half-second pause. A small internal question: is this okay? Then you override it and hit send anyway. Neuroscientists who study decision-making have documented what they call somatic markers where physical signals the body generates ahead of conscious reasoning, based on prior experience with similar situations. The original research, from Antonio Damasio, was about financial risk decisions. But the broader point holds: your nervous system can flag a situation as worth pausing on before your conscious

... LinkedIn posts in your voice. It's web-based, and I used Next.js, I used Better-Auth for authentication, and I added 3 strategies: email/password, Google OAuth and LinkedIn OAuth. I used PostgreSQL as the database and Polar (Stripe alternative) for billing. Other tools I used include: Resend for email, Tailwind CSS for styling, and Antigravity IDE. I hope this helps. Let me know if you need more help. Comment from Viktor Shumylo(@vik_sh): Hi! We re building ... ... because they validated with surveys instead of money. Stack is intentionally boring: Next.js, Supabase (auth + DB), Vercel for hosting, Stripe for billing. The interesting part is the research engine underneath we scrape job boards, Reddit pain threads, and competitor

... with no structure? those are the ones i don't trust Comment from nim(@sup_nim): honestly i think the trust issue comes down to scope. i trust ai way more when its doing one specific thing really well vs trying to be a general assistant that touches evrything. like i built speakeasy which uses ai to convert articles to audio - its just doing text to speech, nothing else. and astrologica generates daily horoscope podcasts from your birth chart ... ... weeks and honestly I m really happy with it, as long as it stays in copilot mode (I review everything before it ships). I ve leaned on it heavily for a new side project I m building ( BenchCanvas ): market + competitor

... here. Startups aren't a race. They're a marathon in an idea maze, and you're far more likely to reach a dead end and kill your startup that way than you are to be crushed by a competitor. Okay, so if launch early and launch often is the 90th percentile advice, there are still cases where it makes sense to stay in stealth later. Also, when you're really, really early, it takes a little bit of time ... ... build a minimum viable or minimum lovable product that you can launch. I'm curious how people at the early stages are thinking about this! If you haven't launched yet or you've consciously

... envied others who could just sleep for 5 hours and still be ok until I read the book 'Why we sleep'. (I might now agree with everything in the book. but that's beside the point) Are you consciously trying to sleep better? Or you already do? How many hours of sleep do you manage to get daily ...