... Apple's App Review Guidelines say you must disclose where personal data goes to third-party AI and get explicit permission first. A privacy policy paragraph doesn't do that. Most of us read Apple's November 2025 guideline update, saw "AI" in it, and filed it under things for the big players. I did too, until I reread it as someone who sends people's journal entries to a language model. Guideline 5.1.2(i) now says: "You must clearly ... ... disclose

... them. The 272K line on the pricing page is a good reminder that the price isn't flat anyway. But the bigger issue is not the bill. Every extra token you send is data you now have to justify, disclose and protect. If your product handles anything personal, the question isn't "can we afford to send this?" It's "does this answer get better enough to be worth sending it?" We learned this the hard way at Murror ...

... list. And there's no small-business exemption. The obligation attaches to whoever makes the platform available to a user, full stop. If you're in scope, California asks for three concrete things, not a vague duty of care. Disclose that the user is talking to AI whenever a reasonable person could mistake it for a human, and for known minors, repeat that disclosure at least every three hours during a session. Maintain protocols for detecting suicidal ideation or self ...

... right and is more useful than either, and the revised version went up this month, so it's worth twenty minutes before you write your Q4 plan. It's NBER working paper 35275, Demirer, Musolff and Yang - "Writing Code vs. Shipping Code." More than 500,000 GitHub developers, matched against their actual AI usage telemetry rather than a survey, in a matched event study design. The headline numbers are large and I have no interest in arguing with them. Cumulative ... ... last eighteen months assumed. Who paid for this, and why I'm citing it anyway. Two of the three authors previously held postdoctoral positions at Microsoft and now work as paid research consultants for the company. That's disclosed

... neither is collapsed into a verdict. What I am genuinely unsure about is whether any of that is enough, or whether a benchmark run by someone selling one of the entries is simply not credible no matter what it discloses, and the only real answer is handing the method to someone with no stake in it. So: if you have run or read one of these, what would you need to see before you believed a number published ...

... keep coming back to the same question with AI phone and operations agents: what should be disclosed before the agent is allowed to do anything that affects another person or system? I do work in this area, so I am biased toward over-communicating. My current instinct is that the minimum should be explicit identity, whether the conversation is being recorded or summarized, what actions the agent is allowed to take, when a human can step in, and some kind ... ... reversible low-risk tasks, but anything that creates a financial commitment, changes identity details, or sends a message the user would not want quoted later should force explicit confirmation. Then the receipt becomes the audit layer: what the agent disclosed

... standard editing. If you deploy emotion recognition or biometric categorisation, you have to tell the people exposed to it. Notably you don't have to explain the purpose just the operation. If you deploy a system generating deepfakes, you disclose. Same for AI-generated text published to inform the public on matters of public interest unless it went through real human review with someone holding editorial responsibility. Spell-check doesn't count. All of it has to be clear, distinguishable ...

... long as it isn t over reactive, I think it could be good! Comment from Ryan Miller(@ryanwrites): @alieksia  I'm worried about false positives too. I would prefer to be able to report something as being AI vs. a system that automatically labels it as AI Comment from Kate Ramakaieva(@kate_ramakaieva): Oh, love it! LinkedIn also introduced "AI Slop" flagging recently, and it is a very logical way to restrict at least part of it! Comment from ... ... even less comments on launches ? Crickets amongst the AI bots? A genuinely good use case is using AI to improve non-English-speaker's contributions. In those cases, it would be lovely if the commenter would self-disclose

... fuzzy on purpose, and fuzzy-on-purpose is exactly where indie products live. If you're in scope, here's what SB 243 actually requires, and I want to make the case that none of it is unreasonable. One: disclose that it's AI when a reasonable person might think they're talking to a human clear and conspicuous, not buried in a terms-of-service page. Two: have a real protocol for self-harm and suicidal ideation ...

... users , because the coverage of this was all about the parts that got delayed . Article 50 of the EU AI Act applies as of 2 August 2026 . In plain terms: An AI system that interacts with people has to disclose that it's AI (unless it's obvious from context) that's your support chatbot, your onboarding assistant. Synthetic media images, audio, video has to be labelled as artificially generated. AI-generated text published to inform the public needs disclosure ...

... actually failed, spanning 22+ countries and industries from fintech to biotech to retail. Each entry includes funding history, investors, a sourced explanation of what actually went wrong, a timeline, and a plain-language lesson, built to be a genuine alternative to expensive tools like Crunchbase that treat failure as a footnote rather than the actual subject. Every entry is manually researched and verified against public reporting, not scraped, and where a figure was never publicly disclosed

... shipped work where 90%+ of the code came from prompting an agent, with me mostly reviewing, testing, and directing the architecture. Never explicitly disclosed that to the client, they just know it got done fast, works well, and stayed in budget. Is that dishonest, or is it no different from using a framework, a library, or a junior dev and not itemizing every line's origin? Nobody used to ask what percentage of a website a developer typed themselves versus

... user and any friends they add. The free app, which offers some paid features, has about 4.3 million monthly active users, according to Sensor Tower data. The previously unreported acquisition closed this spring, Midjourney said. It declined to disclose financial terms. This is big: CEO @banu_guler is now Chief Design Officer at Midjourney. Comment from Alpaysh(@alpaysh): incredible! Comment from Varun S.(@varuns): Wonder what the angle is here. Looking forward to the shifts. Comment from Oluwole Zacheus(@emmadagold ...

... think read is the wrong boundary. In the bridge today there is no privileged read path: `fs_read` and `shell_exec` sit behind the same process-global unlock, and arbitrary shell means read current state can still disclose secrets, hit side-effecting endpoints, acquire locks, or broaden what the model sees. So if I add scoped capabilities, I d model a **bounded observation capability**, not a blanket read capability: same actor/trigger/target/provider, specific observables or version tokens, and it can only ... ... nothing, the reviewer's normal experience becomes an empty screen, and empty screens train people to expect empty. Whether the one that is not empty gets read properly on day forty is an open question. It still beats the alternative

... moved and tmux reattached elsewhere are the same failure underneath: the identity you resolved against stopped being stable, and nothing told you. Which is the argument I have been making all week about session memory generally, so I should disclose the obvious. I hunted Atuin on here today, which is the shell history side of exactly this: history that keeps the directory, the exit code and the duration instead of a flat file that forgets which session it belonged ...

... animation and 8 seconds of silence. Our engineer thought it was a bug when he first saw the spec. "You want me to add... a delay?" Yes. On purpose. Here's why. We ran an experiment: instant AI response vs. an 8-second pause before the same response. The content was identical. The only difference was time. Users who got the pause rated the AI's reflections 40% more helpful. Same words. Same insights. They just felt more considered ... ... honestly surprised us too -- we expected maybe 10-15%. It made us rethink a lot of our other interaction patterns. Comment from Mona Truong(@monatruong_murror): @alieksia   Really sharp question. We haven't run a controlled test on disclosed

... like selling myself either, so I don't. What I do is write comparison pages between the competing products and mine, and put everything into ranking those. The people who land on them are already interested in a competitor, or already using one and unhappy with it. They read what each product does and what it costs, and if mine comes out as the reasonable choice, the conversion happens sooner or later. If you don't have much ... ... broke, what we rebuilt, what surprised us. People root for founders who are honest about the process. Comment from Anmol Dhingra(@anmol_rerato): A rule we landed on for product copy that turned out to apply to promotion too: disclose

... kabandisaikia): I feel Clay is great but gets very expensive for startups. I haven't tried other tools. Comment from Moeez(@moeez1): @sudipta_biswas4  Fair question. From what we ve seen, people don t really react to AI vs human , they react to whether the call is relevant and well-timed. And yes, if someone asks, it s always disclosed

... questions likes that under 10 seconds, end to end, presenting you answer, charts (if needed), all data used and a non-technical explanation what is what (if you drill down to it). When we do it, we don't disclose your data to LLM, while at the same time still using excellent coding agents in realtime to build, run and show you a one-off application to analyze and answer your question, all in under 10 seconds! Under the hood ...

... funding, I had to automate everything, therefore I relied on a cross environment team of AI assistants, that I, painfully and traumatically, intuitively guided to build deterministically. Since patent is pending, I cannot reveal too much, but I have disclosed enough for technical people to understand the architecture. You can also feed my announcement landing page URL to an AI assistant, so it could explain the technology ...

... actual insight, and it's bigger than it looks: making the notification say why it's waiting is what turns a ping into a triage. "Done" makes you tab back to find out what "done" means. "Needs approval vs finished vs

... economics just don't math. A few things I'm chewing on: Is usage-based or credit-based pricing now inevitable for any AI product with an agent layer? Anthropic claims Max 20x gives 20x the usage, but the disclosed numbers work out closer to 6x on Claude Code hours. Are we entering an era where "20x" means tokens, not outcomes? Cursor got dragged for communicating their pricing change poorly. Anthropic seems to be getting ahead of it with ...

... without acknowledging it beforehand is quite disrespectful in my opinion. In addition to trust, decency that comes with being forthright was lost. Definitely not the best way to establish a new relationship. Now, I wait to see if they disclose the recording before I chime in to ask about it at every video call. I think it is a form of etiquette to get consent. Comment from Jared Salois(@jared_salois): Good point! Legality and trust really are two different ... ... they handle client interactions? Comment from Farhad Asbaghipour(@farhad_asbaghipour): I d probably drop immediately trust is gone at that point. It s interesting how many modern workflows are becoming record first, ask later. Curious how others balance efficiency vs

... answer first. If they find it immediately, they cite it. The structure makes extraction easy. Step 2: The Comparison Page Template Comparison pages drove 40% of our citation growth. Here's our exact template: Title: [Brand X] vs [Brand Y]: Which is better for [use case]? Quick verdict (100 words) Who wins and why One sentence summary Feature comparison table Feature Brand X Brand Y Feature A Feature B Feature C Price XX YY Detailed breakdown (by feature) How each ... ... margin of error. Result: This block was cited verbatim by Perplexity within two weeks. Step 5: The 90-Day Update Cycle We analyzed citation decay across 500 pages. The drop-off is steep: Months Since Update Citations (vs

... think we all lie every day (even when we say "Good night" to someone, and we don't even have to mean it sincerely). We also lie on our resumes , we don't fully disclose everything on our tax returns , and we sometimes fake income . And that's what I want to talk about... Cluely CEO Roy Lee admitted that the $7M ARR figure he gave TechCrunch last year was fabricated , calling it the only thing he's "blatantly" lied ...