... people have asked how Mira actually works in practice, so wanted to write this up. The full flow from start to finish: 1. Set up your study (5 minutes) Choose a template Customer Discovery, Concept Testing, UX, Brand Perception, NPS follow-up, and more. Or build your own discussion guide. Mira

... launched Mira on Product Hunt today, and the conversations in the comments have been the most interesting part of the whole day. Researchers asked about probe neutrality whether the follow-up questions an AI asks mid-interview can lead a participant, rather than uncover them. They asked about cultural calibration, whether emotion models trained largely on Western data can accurately read a participant in Jakarta or Nairobi. They asked whether the say/feel mismatch gets surfaced as raw evidence or quietly ... ... before they appear in a report? Proof of regional validation against local human coders? Something structural about how probes are designed? Or is there a ceiling, something you believe only a human moderator can do? Not asking to defend Mira

... three or four moments that actually matter. This observation that the transcript captures what was said but not how or with what underlying feeling is fundamentally what drove us to build the emotion and behavioural layer into Mira. Not to replace researcher interpretation, but to give researchers a way to find the right moments faster. I'm genuinely curious: what's the signal you find yourself wishing you had after you've finished going through a set of interviews? What ... ... review setups that help you capture more of what the transcript loses? Comment from Riya Pariyar(@riya_pariyar): the thing is, transcripts flatten everything into words but a three second pause before answering a question about pricing or a competitor

... moderator catching the right moments. And even the best human moderator misses things, they can't track vocal confidence, facial micro-expressions, and conversation content simultaneously across 40 interviews. At Decode , the reason we built an emotion layer into Mira wasn't to replace researcher judgement. It was specifically to give researchers a way to see the moments where what someone said and what they felt came apart, and to surface those moments as evidence, not as a resolved verdict ... ... many signals simultaneously across so many sessions. The staged easy tasks point is particularly sharp. Building trust early changes participant behaviour for the rest of the session. That's a design principle we've tried to replicate in how Mira

... Founder at @Decode by Entropik here. One of the things that surprises people most about MIRA is that we offer eye tracking: gaze, fixation, dwell time, heatmaps, attention of interest, using a standard consumer webcam. No Tobii. No IR sensors. No specialized hardware of any kind.    The immediate question is always: how?    The honest answer is that webcam-based eye tracking is a fundamentally different problem from hardware-based tracking. Hardware trackers use infrared illumination ... ... biggest advantage here. Getting slightly lower accuracy is often a reasonable tradeoff if it makes the technology much more accessible and easier to use at scale. Comment from Gal Dayan(@galdayan): the honesty about real-world accuracy dropping vs

... making the model's read contestable in the moment: surface the observation ("noticed a pause there anything behind it?") instead of acting on a silent classification. That's your "surfaced not resolved" philosophy applied to the probe itself. Does Mira expose why it probed to the researcher afterward, or just the probe and the response? The audit trail of the classification is where I'd look to catch leading after the fact ...

... guess which element on a busy screen I actually meant, and it edits the wrong one. Then I'm re-prompting, which is its own iteration tax. Curious how much of your rework is this specific thing versus raw model quality. And what's your fix, crop hard, or describe the element in text? Comment from Mustafa Arian(@thenameisarian): @bickov   rough box-and-label works about 70 percent of the time. label has to be specific. not 'header ... ... proxy for design intent. shortcut: if you find yourself describing what should NOT change, that is when to use the sketch. if you only need to point at the one thing to change, box and label. Comment from Mira

... urgent so I was context-switching into three different "modes" a day. What ended up helping wasn't cutting project count, it was a filter that told me which 1-2 things actually needed a decision from me today vs. what could wait. Curious if others here found a similar signal-vs

... indu_thangamuthu): Desktop Comment from Nika(@busmark_w_nika): I do not want to spoil, but... soon, it will be my desktop wallpaper :D Comment from Kseniia Che(@kseniia_che): Downloads, of course. But screenshots are becoming a serious competitor. And then there are my 20 open browser tabs, which seem to worry other people more than they worry me Comment from Suzy Chase(@suzychase): Legit, it's my desktop. It's not even a folder anymore ... ... named things like "report_final_FINAL_v3_useThis.pdf" that I genuinely can't tell apart anymore. The worst part is I know I'll never delete any of them because "what if I need it someday" kicks in every single time. Comment from Mira

... could hold me. Not a hotline. Not a meditation app. Something that understood what grief and mental health actually feel like. I couldn't find it. So I built it. Gentle Hold is a mental health ecosystem - AI support (Mira, always on), real therapists globally, a safe peer community, guided journaling, calming music, and our podcast Elevated Talks. Not an app. An ecosystem. Pioneered in Australia, built for the world. We'd love your support and I'm here ...

... watching the next chapter of tech being written in real time, and even we don't know what the next line is going to be Comment from Marcello Cultrera(@marcello_cultrera1): A similar approach is attempted by Mira Murati at Thinking Machines where most researchers are either quants or mathematicians. Axiom s approach of starting with an AI mathematician is inline with this more recent trend shifting from earlier focus on autoregressive transformers pattern recognition (LLMs/ data science engineering ...

... like to invest in, but that don t have a public offering yet? Which ones? (maybe the dream will come true) For me, it s OpenAI and Stripe. Possibly Product Hunt as the love brand. :) Comment from Elena Mira(@elenamira): I'd buy the "default stack" that compounds over time. Stripe, Vercel, OpenAI, Plaid. Defaults win because switching costs are real. Comment from Nika(@busmark_w_nika): @elenamira  What about Lovable? Why did you choose Vercel over other competitors

... code for people to sign up, as well. Super old school, but better than investing in a set of interviews. I've also tried Reddit searches and scrapes of posts to find pain points that my offering would address, versus testing my idea outright. Sometimes, it's in the positioning rather than the idea itself where you'll gain traction. Comment from Sheraz Abdul Hayee(@sheraz_abdul_hayee): For me what worked was doing the reverse engineering of the existing ... ... share the niche. It is not perfect research but it shows me reactions early and saves time before setting up formal interviews. Comment from Igor Lysenko(@ixord): I would say a little bit from all areas :) Comment from Elena Mira

... everyone! I'm thrilled to join this community and introduce myself. My name is Miras , and I'm 16 y.o founder who is deeply passionate about technology, innovation, and pushing the boundaries of what's possible. Recently, I was incredibly excited to be chosen for the nFactorial Incubator 2025 a summer program for startup creation with a highly competitive 4% acceptance rate for full scholarship. This is where I'm currently working on my passion project, FaceArena. What is FaceArena ...

... need more projects like that Comment from Cristian Stoian Urzica(@cristian_stoian_urzica): @miras808 I like the fact that it is in the fun/games area. I didn't see lot of these projects here on PH Comment from Miras(@miras808): @cristian_stoian_urzica   Thanks so much! I am really passionate about creating truly unique and fun experiences with FaceArena Comment from Miras

... tool like this to quickly validate your idea and get feedback? Please share your thoughts, pain points, or other tools you re using today. Would love to hear your experience! Thanks in advance for any feedback! Mira, Co-Founder Ideastart Comment from Borja DR(@borja_diazroig): Hi Mira

... launches win or lose momentum based on how much effort went into the comment strategy. Curious to hear what others have tested! And if you re launching soon, I'd be happy to swap feedback or upvotes. Comment from Miras(@miras808): This is an incredibly helpful and timely post, thank you for sharing such valuable insights, especially on the importance of the first comment! I'm actually preparing for my Product Hunt launch with FaceArena (my online party game where ... ... grateful if any experienced makers or hunters in this thread would be open to taking a quick look if I share my draft comment here, and perhaps offer any thoughts? Your collective perspective would be highly appreciated! Comment from Miras

... sure why it is happening. And i dont think it is about the actual context of my response, cause If i reply the same exact text as a completely new comment, it gets posted :P weird ! Comment from Elena Mira(@elenamira): @steveb I don t know yet. I need to work with the new design a bit more first to decide. Comment from Luke (@luke_arwayda): It definitely feels somewhat similar to reddit. Will need some time to get used ... ... having the topics on the left side vs

... they are hard to get. Comment from Lawrence(@brainkeeper): I am trying for grants because of the social impact of my product, but I also need some investor money for things not included in the grants. Comment from Elena Mira(@elenamira): @nickanisimov Thanks for the reply, Nick! Why do you consider friends the worst option? Comment from Nick Anisimov(@nickanisimov): The best option is money from clients, then grants, then venture capitalists, and the worst option is friends ...

... OpenAI announced that Sam Altman will be stepping down as the CEO of the company as the board "no longer has confidence in his ability to continue leading OpenAI" The CTO, Mira Murati, will be the interim CEO of OpenAI. What do you think about the future of OpenAI looks like ...