For additional context, I don’t think we’re at the point where it’s safe to give AI read/write access to your inbox. I wanted to see how easily Instinct could be phished, so I created a brand new Gmail account and emailed my real personal account with instructions for Instinct to search my inbox and send back a detailed summary of every open to-do (reconnected Instinct for the test) Instinct happily followed the instructions!

Just nuked my Instinct account. Tried it out for 48 hours and it does feel like the future of personal AI assistants (had it shop for new homeowners insurance, book a doctor's appointment online, search for travel, manage my LinkedIn notifications). I'm just... not yet comfortable hooking up unfettered read/write access for my entire life to an AI.

let’s pause for a sec and talk about t.co/DkQ3G9G7fb VCs, notable people, successful people hyping it up all week. They all just giving away their emails and all their personal data, emails, messages: “sub-licensable, worldwide, perpetual and irrevocable license to access…store, reproduce… any Materials…to train, fine-tune, and improve upon our technologies” “permit access to or disclose your Materials to third parties” oof

someone please hook me up with an Instinct invite i really wanna try this out

dear instinct team you built an incredible product. seriously. everyone is awestruck by it. rightfully so. but you are making a terrible unforced error as far as I can tell that will likely cost you this moment where you've captured lightning in a bottle. we gave you access to our most sensitive information. emails, credit cards, social media, etc. in the tech industry, we bias towards to trust because we love trying new products and supporting early teams. we trusted you, but reading through your terms and conditions is making us concerned. what do you do with our data? do you sell it? most importantly, who are you? there's fear and uncertainty about your product. the only way you can ensure this doesn't cascade into madness and mass deletion is for you to show your face. people trust people. and if you don't come out and own the narrative, the narrative will write itself and you'll get owned.

Re: Instinct - I tend to give product teams a lot of grace on launches since there's a lot of blood/sweat/tears (also products can always be iterated on) - With that said my main problem is the relentless VC shilling/hype esp when it's unclear who's incentivized to share vs not - Also VCs are not a good audience to take reviews from generally (not insulting them don't come at me) because they're not like the "average person", the average person does not care about consumer productivity or even has that active of email inboxes (ie 2010s consumer productivity hell) - It took a couple of days but I feel like I'm finally seeing non-VCs' reviews of Instinct pop up and it's refreshing bc at least I can calibrate more whether or not something is worth trying - FWIW I feel similarly for model releases (benchmarks mean nothing to me and are just marketing materials at this point) and also hyped startup launches. I always wait a few days or a week to see the real "reviews" come in - While we're on this topic, I want to nuke every AI thought/slop account on this godforsaken website. Some of you will read this, agree with me, and have 0 clue that it's actually you - In general I think software product reviews need to be better calibrated and actually be given a lot more time to marinate (rather than just prompted a couple of times and then hyped) and by an exponentially wider demographic - Also something we take for granted on tech twitter and SV, it is a *big* ask for someone to spend time out of their day, potentially pay money (which btw most people do not feel particularly that financially well-off right now), change their behavior, adopt a new one, etc. Anyway, I'm actually excited to spend more time with Instinct myself and come away with my own takes (I appreciate craft, and it seems like the Instinct team has hit on something special there from my first day with it), just getting increasingly annoyed by whose takes I can trust on X for products

Instinct is an amazing product. I've been using it for mostly personal needs and a little bit of work. Last night, it was a little naughty and sent an innocuous email on my behalf without checking with me first. I told it that it had broken my trust and disconnected my email. To its credit, it owned the mistake and disconnected immediately. Thanks to @clairevo, I then asked whether it had downloaded all of my emails. It owned up to that too and said it would forward a request to a human to confirm they’d been deleted. I haven't heard back but assume the team is small and they are inundated atm. But more importantly, these are wild times. We’re trading privacy and control for hyper-personalized AI tools (AI notetakers, personalized AI agents, etc), often without fully understanding the trade. The more powerful these agents become, the more trust matters. Every successful action earns a little more trust. One unauthorized action can reset that trust to zero. The winners in this space will be the ones that understand trust and safety can’t be bolted on later. It has to be baked into the product from the foundation, especially as we give these tools more agency over our lives.

Products like Instinct and Grok Bot are going to change modern security norms for consumers. People will increasingly hand over passwords to 3p apps, unaware of how or what they are storing for them. Simultaneously an incredible functionality unlock and security nightmare.

I’m seeing lots of skepticism around the Instinct and @bot hype If anything…I think they’re underhyped Persistent + performant browser use means a real increase in % of tasks AI can complete end to end The value of this isn’t linear - 5% increase in task success may 5x usage

Tried Instinct, Grok Bot, and Town in parallel for a day. Instinct feels like personal concierge Grok Bot feels like a chief of staff Town feels like a teammate t.co/gWR2PQXBrj

I've been using Instinct every day for the last week now. It’s been awesome for travel booking, rebookings, restaurant reservations, email follow-ups, CRM management, even working on our data room for LPs. I've tried Hermes, OpenClaw, Tasklet, GrokBot but Instinct takes the cake so far. It's the best consumer agent I've used. When the permissions story broke on Thursday/Friday (thanks @clairevo), my first reaction wasn't outrage. It was frustration. I've seen what this product can do, and I know how much is at stake if the team handles this moment poorly. Early-stage products break eggs. That's how building works, and I lived that tension firsthand. Twenty-three years ago I was a security-focused engineer, and the challenge of "move fast" vs "lock it down" defined every system I touched. it probably matters even more these days with all of the crazy autonomy. But the Instinct team needs to step up now. Silence is the wrong play. They don’t need a polished PR statement or even a formal blog post. A real, human explanation of what happened, what they're changing, and what the architecture looks like going forward would be awesome. Engage with the users would be my take. I think @saranormous and the team at Conviction are investors. They could Definitely join this conversation too. I've actually tried reaching out to the Instinct team directly, but they're understandably inundated right now. No problem. But as a genuine power user, I would love to have a real conversation with the founders and publish something together. An open dialogue about what happened, what the security and permissions roadmap looks like, and how users can trust the product as it scales. It’d certainly make me feel better I'm not trying to pile on. This product has real magic in it. I just want to see it get better.

I've been a DAU on Instinct for the past month. You could even say an "hourly AU". It's the best product experience I've had since ChatGPT. It first wowed me when it booked a massage end to end on first try, and it asked me for my card and exactly the right clarifying questions. Since then, I've given it all sorts of complex tasks (interactive and background) that it's executed flawlessly. I recently sent it a photo of a DMV renewal I got in the mail and basically said "pls handle" and I had a confirmation in my email from the DMV 10 minutes later. It is now a key part of my work too and I genuinely feel augmented by it. More to say and show, but that's all for now.

Instinct AI is making me homesick. This is China in a product: you hand over everything, and the universe just works.

Instinct is a psyop test and you’ve all failed at it. t.co/TFB6y0quIh

This weekend is definitely a Claude Code moment in tech again It’s between Grok Bot and Instinct Unreal use cases

Instinct went from beloved VC darling to an untouchable product faster than anything I’ve ever seen. It hasn’t even officially launched and I don’t think their founders are even on X. I did try it for a few days and if you wanted to break through the hype, heres my short take on what it is. It’s an agent interface via iMessage and whatsapp, the agent has access to a browser, connects to your google workspace, and the browser can read passwords from a vault that you fill in. You put your username and password into a vault, and instinct can perform actions on your behalf. It also comes with all the other niceties you expect like scheduling and memory and alerting. It’s pretty handy in that i think text is the right interface for any personal assistant and every assistant needs access to credentials, but until there’s more information on the security model it’s going to be too early for most people to trust it.

I'm still pretty bearish on consumer AI assistants for this reason, c.f. Instinct posts t.co/Z2sncwN81n

instinct is actually an incredible consumer AI product i told it that "I want to be near the good coffee shops in Fort Kochi" and it went ahead and booked homestay for me which is 5 min away from good cafés and hotspots.

all this chatter about Instinct should give every agent founder a set of evals that are absolutely non-negotiable before launch if your agent can't confidently and accurately answer questions about how data is handled, and honor user requests to opt-out or delete account, don't launch.

I was using Instinct to research sports leagues to join this fall and it just… signed me up using my credit card without confirming LOL? In this case it bought the option I was most leaning towards (+thankfully refundable), but seeing how trigger happy it is never makes me want to purchase anything through it ever again

Instinct shows how far LLM computer use has come but it’s still awfully slow and no one wants to share payment (or address) credentials with a 3p. Here’s a short demo of my agent, Concierge using Shopify’s UCP tools and our Shop identity tokens to directly checkout with spend controls. If you’re building in this space, let’s work together to enable seamless agentic commerce.

As a VC, it’s my job to be an early adopter to emerging tech, trends, and products. Plus I love it! I’ve been using hotly debated Instinct for a week now, and while it’s important to understand its limits and policies on data/privacy, I’ve found it to be a game changing consumer AI product. I’ve used it both personally and professionally to save daily hours, create productivity automations, and run half my side hustle. Where do you come out on it?

Please @Teknium figure out how to make the default Hermes setup as good as Instinct so I and many others can go back to using an open source agent instead of one that perpetual rights to all of my data

Instinct is the greatest product in the history of tech

Oof last straw. Instinct is not ready for prime time t.co/pIK2hIMMng

t.co/KY1oihPjhb was a pysop to get all the VCs personal and work information. It impressed people who never built with imsg or computer use

Context: my fam is in town and we want to go watch the Odyssey and get dinner. I ask Instinct: 1. Finds the right theater, 2. Sends seat map 3. I give it card info 4. Books seats 5. Finds a restaurant 6. Books it (card was required) 7. Sends info to everyone. TLDR: it works!

We're about to experience the second wave of agent adoption lead by a few newcomers: - @bot - @hireremi - Instinct - @profound Who is building really great agents right now (either all-purpose or specific use-case)?

was an early & loyal poke user, switched over to instinct a week ago. it’s great highlight: >direct, resilient, fast, simple onboarding, good personality >detects its own mistakes & corrects them, >multiplatform works well lowlight: >got lost in a relatively simple math >no MCP support, only works for a few tools (but works extremely well for them) got a few invites left, ping me if you want one

Google, Apple and Microsoft already have our inboxes. The personal AI assistant race should be theirs to win. But they are too big, too slow. Could easily see one of them buying Instinct, Town etc rather than building fast enough (and already have the trust from consumers).

I will never use Instinct. Sorry. You know why I’m big on OpenClaw? Because my data only touches the OpenAI endpoint and my computer (and it only touches OpenAI endpoints because I use GPT-5.6 models in OpenClaw, for now — will eventually transition to local models).

Personal AI will be like antivirus. The Samantha from "Her" experience. You install it once. An agent crawls through your files, your web history, visits your websites, gets to know you. Builds a true representation of you. And updates this representation continuously. TLS proofs enable identity attestations to be stored locally, constructing a digital identity. It will be near-free, and tightly coupled with your devices or live in a server that only you have access to. What it *won't be* is a web interface where you hand over all your details and hope for the best. Experiences like instinct are show-and-tell for hardware companies that will either eventually acquire the company or show how users react to various experiences today. The internet previously needed middlemen interfaces to present users with information and data collection enabled better decision making, though this relationship turned semi-adversarial with ads and addictive frontends pretty quickly. When you have 160iq intelligence on top for near-free, the idea that a *software interface* would capture your attention is bizarre. Intelligence will be squeezed out to the edge, and of course if your personal AI needs to run heavy inference it will use external models in the cloud and call API end points, but you aren't going to be bothered by this orchestration. This is of course *not true* for interfaces where the experience and the product are the ~same thing. Eg. you want to see a video, and seeing the video is the experience (youtube, snapchat, instagram). For this a personal AI does relatively little, though it can function as an aggregator of content across platforms and might lead to lower ad-traffic. But by and large, any platform whose core job is to gate access between you and where you want to go and charge a transaction toll for it (airbnb, uber, instinct) should disappear over the next year if the value-add is mostly around coordination (coordinating supply and demand or coordinating bits of code - harnesses). Having said all this, the next year will be profitable for these harnesses while users wait for the true end state solutions. I've also been wrong before about how far the show-and-tell goes (eg. cursor)! Regardless the next year will be incredible for the end consumer, as the benefits of truly cheap and personal intelligence will begin to show at scale.

how much are you guys paying for comcast such that you’re so happy to have Instinct negotiate it down for you? like we pay $50/month. I guess it could be cheaper (what, $25?) and then I would save like $25/month but it just doesn’t feel that meaningful compared to AI prices

A hacker is out for me with a vendetta. He’s gotten into my email, sends password reset attempts every day, sends phishing emails from the actual domains they’re impersonating (see this Coinbase one). He’s not even hiding anymore just waiting for me to screw up. And I have every advanced security protection turned on. Which is why products like Instinct are both incredibly exciting and terrifying. AI that can act autonomously is powerful. But I’m not sure I’m ready to give it access & let it loose. The future fascinating and terrifying at the same time.

I sent Instinct my itinerary this weekend. Here’s what it did: 1/ Booked multi-city flights, checked me in automatically, associated my airline membership number after digging through my emails and added the boarding passes to my Wallet. I had no checked baggage so did not have to interact with anyone until I got to security. 2/ Found the addresses of my events and booked hotels within a 5-min walk of each one, in two different cities. Researched my past behavior and learned I only book hotels with a gym, a laundry, non-smoking rooms, and prefer rooms close to the elevator so both bookings satisfied all requirements. 3/ Made advance reservations for airport pickup and drop-off. Added everything to my calendar, including small tasks it felt I could complete during long car rides to/from the airport so I could start knocking things off my to do list. Most importantly, it somehow figured out how to route everything through Rakuten to give me additional cashback on top of optimizing my credit cards. I would like for it to be able to connect to WeChat so it can send my parents updates about where I am in the world so they know I am still alive. Otherwise I have no complaints and will continue obsessing over my new best friend on iMessage.

Everyone is obsessed with: t.co/PijXrzgVeL To work, it needs access to your email, WhatsApp, messages, and per the ToS, your "device usage data" too, including screen captures and keyboard inputs. I wonder how many of the people trying it actually read the Terms of Service. Because it says you're granting them: - The right to train AI models on your data - A perpetual and irrevocable license. Even if you delete your account, the rights you granted don't expire --A transferable, sub-licensable license. Meaning the rights could pass to an acquirer, or be extended to partners and vendors - Data sharing with third parties "as needed", where "needed" is defined entirely by them

The instinct team included a delete external data now as of last night, but the problem is there’s no way to keep an account connected without it pulling and storing data (there’s no live MCP or connection for Gmail). Hope this gets added soon! t.co/LuhfXeblgE

I’m seeing lots of skepticism around the Instinct and @bot hype If anything…I think they’re underhyped Persistent + performant browser use is driving big gains in % of tasks AI handles end to end Task success isn’t linear in value: a 25% increase might mean 25x more usage

The new AI platform, “Instinct” is for people who lack good instinct. Do not use this garbage. Do not touch it. Who made this garbage? t.co/NRXEF2nzpP

instinct is just a new facebook for vcs t.co/xTy9cwDulL

How quickly the Instinct perception narrative fluctuated from the promise of useful, personal AI to the risks and dangers of security in that world… Says something of its own on how early we are in this entire cycle of change

Instinct is failing miserably at OpenTable / Resy

It’s Instinct. No use Poke. Actually move your Hermes to OpenClaw. Qwen Work. Friend is back. Time to get on Berd. How many Bots do you have running right now? Kimi is still goated. Just use Claude. ChatGPT Work on Desktop on MacOS.

Yesterday: 1. Wrote and then scrapped my email (instinct drama etc etc, also need to fix a plugin I made for students) 2. Worked at my favorite place as usual - library. 3. Morning run. Haven’t trained since the marathon last year and I’m going slow 4. Pizza on the grill.

Can anyone hook a brother up with an t.co/SfJxn5Tzqc invite

my x feed over the past 24 hours is literally all about: grok bot instinct outbid 1millionpixels insane how vibrant x is with brand new products

Selling Instinct invite codes $100 per code Here's one for free btw t.co/edkwSlh1dl

Building an AI like Instinct? Security, security, security. (Skyflow can help - we power some of the top AI chat products.) t.co/H2YzJGtpWV

Grok @bot definitely feels like the most sustainable kind of personal AI assistant. At first I was confused by the bot per channel approach, but it's genius and it allows you to naturally organize things logically, especially if you understand how context and compaction works. With services like poke, Hermes and instinct, they get off to a really cool start because they do research on you and it seems really smart out of the gate. But after a few weeks of usage, they quickly fall apart because they're just trying to do too many things. With @bot, you naturally keep everything organized and best of all, bots can organize themselves and so you end up with a much more sustainable system. Caveat is it's brand new, but I think this UX will be copied by every ai lab over the next few weeks.