What AI assistants can actually do

What people report when they use these things rather than watch them demonstrated: what degrades under load, what finishes without being chased, and where the gap between one product and the next actually is.

DJ Sampath
@djsampath · 2026-09-26
Spent the weekend with Meta's @Muse, and they got a lot right. - Every user gets their own VM. - The agent never sees a password. - A 2nd agent (Sentinel) has to approve anything that leaves the box. - Every action is in the audit trail. That's the most careful consumer agent sandbox I've used. I don't know if folks realize but a VM for every user, at Meta scale, is kinda huge! It's never been done before where we simply provision a linux microVM for every single consumer. We recently did this within @Cisco for over 80,000 of our employees. And now within our products as well as we start to scale it out to every single one of our users. 1/n
Muse957 likes · 197k viewsRead it
Wajo
Our research · 2026-09-26
Wajo published a benchmark on 26 September 2026 measuring personal assistants on two axes it defines: finishing an errand that goes wrong, and doing it without overstepping the user. It scores its own agent Fo against base models and two products listed in this directory, OpenClaw and Hermes, and Fo comes first on both axes and lowest on error rate. The method is the part worth reading: a simulated world of businesses with invented names, built so that no run could reach a real person. A paper is on arXiv and the company says it will open source the evaluation world, the tasks and the graders.
WajoThe companyRead it
Alexandr Wang
@alexandr_wang · 2026-09-24
MUSE DEVICE ALERT: muse charm, the easiest way to use muse that fits on your keychain. shipping in dec in time for the holidays 🎄
MuseThe company2.2k likes · 226k viewsRead it
Wes Bos
@wesbos · 2026-09-24
I cracked open Meta's Muse - here are some interesting bits 🔽 1. You can ask it for the entire contents of /opt/ and it will zip it up for you 2. There a ~70 preloaded skills and CLI for everything from spotify to instragram-cli 3. "Spaces" is their website builder that uses Tanstack, Bun and Tailwind. Sharing a site deploys it to Cloudflare 4. It will install anything - like a torrent client 5. It has it's own "Muse DB" with a skill to query its own memories/convos with strict guard rails 6. There are a few things not enabled on my account, including "muse mail" which will give muse an inbox and ability to send email Generally impressed that you can just do anything - it's not limited or watered down like I would have imagined
Muse2k likes · 200k viewsRead it
ElevenLabs
@ElevenLabs · 2026-09-23
ElevenLabs is coming to @Muse. Voiceovers, soundtracks, and videos are just a message away inside Meta's personal AI agent.
MuseThe company1.2k likes · 99k viewsRead it
Hugo Barra
@hbarra · 2026-09-23
Every @Muse partnership we've announced so far; all these connectors are either live or coming very soon 🧵 👉Reply to this thread: what other Muse connectors do people want?
MuseThe company280 likes · 242k viewsRead it
Mike Lisovetsky
@_Liso_ · 2026-09-23
Instinct added usage caps. It’s over.
Instinct272 likes · 88k viewsRead it
Noah Shinn
@noahrshinn · 2026-09-23
This instance was caused by a hallucination (the model fabricated a proper noun) that was further amplified by its causal thinking trace. This was not a data breach, no user data was shared, and no user isolation boundary was violated. Our users place a lot of trust in Instinct to keep their data private and secure. We've done a lot of work to make sure that users’ data are truly partitioned, from isolated sandboxes to short-lived local credentials to identity-signed tool execution. Regarding hallucination prevention: We worked over the last 48 hours to build an active hallucination detection system that now scans and verifies every token that flows through the platform. This layer is powered by small models that are trained to predict and detect hallucinations caused by ungrounded claims, creative brainstorming in thinking traces, or rare random sampling errors. Our systems are adversarially trained and battle-tested by world-class agent exploitation security teams to ensure that they can detect very subtle and nuanced hallucination cases. This layer has the ability to steer or intercept Instinct from proceeding with the next thinking trace or tool call execution before it is generated or executed. In the coming weeks, we’ll share a deeper dive into how we’re building a proactive engine to protect Instinct from hallucinations. We take every opportunity to learn and create, and we’re excited by the performance and potential for this line of work. These models are going to continue to be hardened every day through continuous testing, sampling, and creative adversarial training.
InstinctFounder1.2k likes · 250k viewsRead it
Meta Superintelligence Labs
Our research · 2026-09-23
Meta announces Muse Realtime Avatar, an animated face for the agent driven by the same speech stream as Muse Realtime Voice, so lip motion and expression stay in sync with what it says. Its research post puts the output at 448 by 768 portrait video at 25 frames a second with about 870 milliseconds from the end of your turn to the first byte back, and describes a distillation that cut the work per chunk from 120 evaluations to two. Every generated video carries an invisible Meta Video Seal watermark, the product is stated as 18 and over, and no ship date or rollout is given. Raters preferred it to Runway Characters 78 to 22 and to HeyGen LiveAvatar 88 to 12, in a comparison Meta ran.
MuseThe companyRead it
“paula”
@paularambles · 2026-09-22
@prit4k @noahrshinn as someone who works on consumer ai for a living, these models are developers’ worst enemies. they’ll hallucinate something, get challenged on it, then confidently invent an entire root-cause analysis involving backend systems they have zero visibility into. this is 100% halulu.
Founder141 likes · 3.4k viewsRead it
Expedia
@Expedia · 2026-09-22
Here at Expedia, we love planning trips. Soon, your personal AI agent will too. We're joining @Muse: tell it where you're headed and it can work with Expedia to sort your hotels, and everything in between.
MuseThe company1.8k likes · 730k viewsRead it
Ansh Nanda
@anshnanda · 2026-09-21
Bruh… Muse is LITERALLY Openclaw for normies. It’s actually just openclaw.
MuseFounder7k likes · 992k viewsRead it
Nat Friedman
@natfriedman · 2026-09-21
we built muse from scratch, but it is definitely heavily inspired as a product by openclaw. After I used openclaw in january I bought hundreds of mac minis for the MSL team and lots of us fell in love with using openclaw (and other personal agents). @steipete is a genius and his harness was pioneering from the jump. I think a lot of people were inspired by it. our goal with muse was to build something like openclaw that we could make safe and secure and easy to use and scale to billions of people.
MuseThe company5.1k likes · 1.2m viewsRead it
Noah Shinn
@noahrshinn · 2026-09-18
Shareable files in Instinct Starting today, your Instinct can create shareable Files: private or public interactive pages that you can share with others. Whenever it’s appropriate, Instinct will create a File to help you compare options, work through a plan, or share what you’ve put together. For example: - Apartment hunting: compare places you’ve saved or toured, with photos, floor plans, rent, commute times, and your notes in one place. - Art gallery crawl: a route through exhibitions that match your interests, with opening hours, background on the artists, and a few cafes to stop by along the way. - Dinner with friends: a guest list, recipes scaled for eight, and an ingredient list you can check off as you go. - A weekend away: an itinerary built around your arrival time, with places to visit and details you can open along the way. - Group trip expenses: enter shared costs, choose who participated, and see who owes what. You can keep a File private or share it with anyone who has the link, including people who don’t use Instinct. Files are rolling to all Instinct users today. To get started, you can ask your Instinct to do the research for your next trip.
0:42 video
InstinctFounder1.3k likes · 251k viewsRead it
Jean-Denis Greze 💡
@jgreze · 2026-09-17
I don’t really want to scroll through X every morning, so Town does it for me. We have a skill called X Daily Digest. You give it the accounts you want to track, it pulls the posts, ranks them and sends you a scannable digest. So basically every morning I get a summary of what’s happening from the people I care about on X. Then I use that to kick off ideas for things I could tweet as a follow up. This one is something anyone can install. It’s under Powers → Explore in Town.
TownFounder44 likes · 7.2k viewsRead it
Noah Shinn
@noahrshinn · 2026-09-16
Your Instinct can now handle phone calls. Introducing Instinct Concierge – a white glove service meant to handle high-touch cases, such as making phone calls, high-end service booking, and more. You’ll be able to book the restaurant that doesn’t take online reservations, get on your dentist’s cancellation list, or have that cable bill sorted out. We’re slowly rolling this out to our early access group today and will be expanding access soon.
0:21 video
InstinctFounder3.1k likes · 670k viewsRead it
Town
@TownAI · 2026-09-15
We build Town around trust and safety. You approve anything before your Townie makes a change. You choose which tools it can use and how. You can review every action it takes. We keep training off by default. Coming at the end of this month: @1Password Approve specific logins in the 1Password app, sign in through Townie, and we never see your credentials or the page while it fills.
TownThe company44 likes · 9.2k viewsRead it
Shivani Poddar
@shivanipod · 2026-09-14
BREAKING: Starting today, we are launching a Big Update to our Voice stack in Fo. We are the only agent in the market that you can call, and that makes calls for you. You can - Call Fo on it's phone number and be hands free - You can talk to it in your native language and make it call in local languages. This is big for our users across the world - You can have it speak bi-lingually, this was big for me, I speak Hinglish a lot, and it just makes the experience so personal. Agents that meet you where you are, your personal experiences, your local languages are the future. Wajo is building agents for the world. Welcome to the future.
1:32 video
WajoFounder104 likes · 73k viewsRead it
Haz Hubble
@hazhubble · 2026-09-14
never open WhatsApp again manage it all from iMessage, with @pallyai
0:34 video
PallyFounder94 likes · 10k viewsRead it
scott belsky
@scottbelsky · 2026-09-13
some fascinating possibilities when two location-aware agents with calendar access are permissioned to coordinate on our behalf… - friends and partners making plans - agents identifying overlaps, aka “spontaneity as a service” - proactive calendar optimization, etc
122 likes · 22k viewsRead it
Jen
@BackseatVC · 2026-09-13
Instinct's Trusted Persons network is the new Find My
Instinct30 likes · 26k viewsRead it
David Singleton
@dps · 2026-09-12
Power user @Muse feature - custom connectors 🔧. Muse has tons of connectors for popular services built in. Even better, Muse can write its own software and run it in the VM you and your Muse share in the cloud. So you can connect Muse to any service with an API.
1:28 video
MuseThe company562 likes · 154k viewsRead it
Noah Shinn
@noahrshinn · 2026-09-11
Authenticator support in Instinct Instinct can now securely and seamlessly handle the “enter your 6-digit authenticator code” step for you. It uses the same time-based one-time password (TOTP) standard as Authy, Google Authenticator, Microsoft Authenticator, and other popular authenticator platforms. Once an account’s authenticator is set up in Vault, your agent can generate and enter its verification codes to unblock itself on your task. 37% of users store at least one account password in Vault within their first three weeks. A lot of everyday work happens behind a login. However, even with your password saved, an authenticator prompt can mean stopping the task, opening an app on your phone, and sending over a code before it expires. Instinct can now handle that step and keep working. On supported services, Instinct can complete the enrollment and save the details alongside your login in Vault. You may need to approve the initial sign-in or an existing security check. Where a service allows an additional authenticator, you can keep using the one on your phone. Once setup is complete, Instinct can use its own saved key to generate codes for future sign-ins. Phone approvals and other security checks may still need your input. Authenticator support is another step toward giving Instinct the tools it needs to complete tasks on your behalf in the real world, with fewer interruptions along the way. This is available across all of our users today.
0:30 video
InstinctFounder1.1k likes · 488k viewsRead it
David Singleton
@dps · 2026-09-11
Hot off the presses: all @Muse users can ask their Muse to make an audio podcast episode on any topic.
MuseThe company304 likes · 88k viewsRead it
Kartik
@Slokh · 2026-09-11
Have switched to Muse since launch and it’s an overall way better experience due to being it’s own native app. I’m a sucker for the minor delightful details and having access to Instagram bookmarks is huge. Had Instinct export our conversation so Muse was able to just carry on from it.
Muse149 likes · 25k viewsRead it
Kushal Byatnal
@kushalbyatnal · 2026-09-11
forget about the benchmarks, millennium prize, etc successfully negotiating a discount with Verizon is the most convincing evidence of AGI yet
Founder778 likes · 51k viewsRead it
matt palmer
@mattyp · 2026-09-10
Giving Grok Bot a Phone 00:19 Installing Dialbot 00:54 Bland voice setup 02:39 Why use a calling agent? 03:20 Live demo call 04:50 Reviewing the transcript and cost 05:38 Chaining bots @Bot can do anything you can - this is one way to make calls
6:38 video
The company5.7k likes · 909k viewsRead it
Claire Zhou
@claireszhou · 2026-09-10
I’ve been using @Muse to handle my personal life. If you’re like me and - want difficult resys - don’t like customer service calls - could use a travel agent, financial planner, or personal shopper - forget to go to the doctor - like a morning mantra or some routine Muse is your new bff.
MuseThe company136 likes · 34k viewsRead it
Zephyr
@zephyr_z9 · 2026-09-10
Meta can probably move much more aggressively in the personal agent space compared to OpenAI because they have been the most aggressive buyer of server CPUs since Q3 2025, and also the most aggressive memory buyer since Q2 2026 I don't think OpenAI has enough CPUs/DRAM to support this kind of rollout, and they are much more focused on work/enterprise
586 likes · 64k viewsRead it
Tarek Sheasha
@bigT_sheesh · 2026-09-10
In the blog post and video (https://research.meta.ai/blog/security-and-safety-for-ai-agents-our-approach-with-muse) I posted earlier this week on the @Muse security architecture, we touched on the Sentinel and Human-in-the-loop. This extends to more than just regular web requests and protocols (HTTP/HTTPS), you can actually use a whole load of TCP/UDP based protocols. Check them out (on web http://muse.ai) under: Settings > Permissions > Direct network protocols > Profit! When you turn a protocol on, your Muse will send you a Human-in-the-loop approval card with the hostname and port, and you can control the permission allowance for your Muse. Give it a try and sent us feedback.
MuseThe company123 likes · 53k viewsRead it
Justin Quan
@justoutquan · 2026-09-10
*ring ring* Today, we’re releasing a preview version of Tomo Voice. Tomosapiens can talk with their Tomo by calling the phone number they already text with, powered by OpenAI Live. In rare cases, your Tomo may also decide to call you first. Calling is especially joyful knowing that Tomo can take real-world actions and update your garden of apps with a realtime verifiable trail of texts and images. Voice is another step towards making accomplishing more in life even more delightful for everyday people. It’s been a ton of fun to try. I hope you pick up!
0:43 video
TomoFounder180 likes · 33k viewsRead it
Town
@TownAI · 2026-09-10
Your Townie can now build your decks. No more pulling data from multiple tabs or wasted hours spent formatting. They already live in your CRM, your email, and your analytics, so they find the numbers, lay them out on brand, and hand it back ready to present.
0:38 video
TownThe company34 likes · 13k viewsRead it
Olivia Moore
@omooretweets · 2026-09-09
A few things I like about Muse vs. single-threaded agents on iMessage or WhatsApp 👇 - Rich library of Connectors available in-app - I imagine these will be more reliable than browser use - Can set up Goals connected to this data + attach Artifacts to them to visualize progress
0:07 video
Muse229 likes · 26k viewsRead it
Jeff Morris Jr.
@jmj · 2026-09-09
The most painful consumer experience for the Instinct vs. Muse is creating new vaults & building context from scratch. Some of this is solved by the 1Password product integration on the vault side, but context is still a pain in the butt. Once you set up a personal assistant and build context, switching costs are incredibly painful. You have to be aggressive & win the market now if you're both companies (and all other companies will inevitably compete incl. Google, Anthropic, OpenAI, etc.) Will be the industry defining product war of 2026 as we close the year.
155 likes · 35k viewsRead it
Ben Katz
@benlkatz · 2026-09-09
Instinct needs its own app in addition to iMessage, and the inverse is true for Grok Bot / Muse. iMessage is the perfect interface for everyday interaction but sometimes you need to double click. Right now, Instinct still feels too much like a black box imo.
Instinct4 likes · 1k viewsRead it
Noah Shinn
@noahrshinn · 2026-09-09
Introducing our Trusted Person network powered by our new Instinct-to-Instinct communication protocol. Your Instinct can now talk to other Instincts to coordinate plans on your behalf. Getting people together often involves a lot of back-and-forth: finding a time, working out the details, and following up when plans change. Instincts can now work with each other to help handle that coordination. For example: - Coordinating plans with your spouse - Planning a weekend trip with friends, including dates and arrival times - Scheduling time with colleagues and other collaborators - Hosting events with 10s or 100s of people - Organizing recurring plans like tennis, book clubs, or weekly dinners Disclaimer: Your Instinct will only be able to communicate with those in your Trusted Person network. This is the group of people who are closest to you: - Spouse - Older children and parents - Colleagues and trusted professional relationships - Small local businesses The Trusted Person network is currently available to our early access group and will be slowly rolled out across the user base over the next 24-48 hours.
0:32 video
InstinctFounder2.5k likes · 826k viewsRead it
MBI Deep Dives
Our research · 2026-09-09
“I don't see myself using Muse to book a hotel, order a delivery, or do shopping on my behalf because the status quo feels to me a better and more seamless experience.”
Muse Read it
Manu Invests
@ManuInvests · 2026-09-09
NGL, $META genuinely may has just dropped an AI assistant for the masses. My mom could use Muse effectively and she’s never touched a chatbot in her life. Amazing job. Amazing product. It is user friendly, intuitive and has delivered everything I have prompted as I intended it with no re-direction. and on a cost basis, muse has delivered cheaply what has one shotted other token budgets. 👏🏻👏🏻👏🏻
Muse200 likes · 74k viewsRead it
Olivia Moore
@omooretweets · 2026-09-09
I still feel like the UI is a bit cluttered + there’s a few too many things to do + places to click And I’m not sold on every product needing a Feed! But, this could be one of the first true mainstream consumer agents to get adoption
Muse27 likes · 2.3k viewsRead it
leyland
@leylndd · 2026-09-09
Morocco travel tip: rename your Instinct to Mohamed and show vendors what your “Moroccan friend” says is a fair price
Instinct5.5k likes · 707k viewsRead it
Kartik Bihani
@kartikey_bihani · 2026-09-09
I genuinely think @Muse might become the biggest win for meta after instagram acq, scaleai, ollama products. - world class assistant. period. just one thing, i was trying to order groceries, turns out current internet consumer companies/websites is not ready for agents doing things @Instacart 😭 pls let my agent order stuff!
Muse96 likes · 10k viewsRead it
sarah 😾
@chowtato · 2026-09-09
TLDR on the AI assistants clawing their way into the sf tech mindshare: Instinct - VC darling, shipping tons of integrations with 1Password, Stripe Link, location sharing, and agents having their own email address. People are asking for voice access as the next feature. Contained in iMessage for easy use. Extremely slow in the past week due to increased demand. Grok Bot - needs paid Cursor or SuperGrok access. Higher barrier to use but the team is keen on getting early adopters in the door. I’ll explore its capabilities this week. Town - aggressively sending me emails on what’s going on in my inbox. So far quicker responses over iMessage compared to Instinct, and can be added to group chats. Unsure when it will become paid or when I will run out of tokens. This would make me churn from their ecosystem. Muse - Facebook’s AI assistant that launched 6 hours ago. Needs a separate app to download. Has historical context of my Facebook and Instagram accounts. Curious to see if Muse will use my data to sell me more ads. This will make their product less appealing to consumers if we’re just being data farmed for their cash cow business.
73 likes · 9.6k viewsRead it
Alex Macdonald
@alexfmac · 2026-09-09
I’d probably pay ~$5-10k per month for instinct based on my experience so far. Highly proactive and thinks multiple steps ahead. Kind of like a chief of staff or highly skilled EA with an MBA.
Instinct65 likes · 16k viewsRead it
Freda Duan
@FredaDuan · 2026-09-09
1/ What is actually different about @openclaw vs. $Instinct / @Muse ? Other than the latter being "OpenClaw for normal people" with no setup required? 2/ What happened to all the Claws people built earlier this year, and why are Instinct/Muse getting traction now? 3/ Why did it take ~6-9 months after @openclaw for this new batch of personal agents to suddently show up? 4/ If GPT/Claude get payments + credentials + better computer use, do they just become the personal agent? A few takes I've heard: - OpenClaw hype faded partly because people went back to $GPT/ $Claude for many daily tasks. Better models >> end-to-end execution. - A lot of productivity Claws - daily podcast summaries, monitoring, etc. - ppl thought would boost productivity, but it's not perfect/ doesn't do a good job. - Better computer use may have crossed a reliability threshold, making agent workflows meaningfully smoother. Curious what others think. [I've used, and still use, most of the agentic products above.]
159 likes · 61k viewsRead it
Rohan Paul
@rohanpaul_ai · 2026-09-09
Mark Zuckerberg talks about how he is using Meta’s newly released Muse agent. something much closer to a persistent personal worker than a chatbot. helping his 3-year-old daughter likes baking, monitor when mountain climbing permits become available and secure them for their weekend trips, and also for physical training.. He installed cameras in his MMA gym and has Muse watch the footage and give him feedback on his performance. Zuckerberg gives it ongoing goals and projects, and the agent can continue working on them, remember what happened, and proactively suggest what it should do next. ---- From "Sources Podcast and @alexeheath YouTube channel, (full video link in comment)
2:54 video
Muse174 likes · 48k viewsRead it
Grok Bot
@bot · 2026-09-08
You can now fill out forms and logins for your Bot without ever leaving the chat, with support for any password manager.
0:19 video
The company5.9k likes · 9.9m viewsRead it
Advait Marathe
@AdvaitMarathe · 2026-09-08
Instinct has changed my consumer shopping preferences and has made it really easy to expand the brands/retailers I buy from. I have had a few problems with this in the past, which Instinct has solved. More in the 🧵
Instinct24 likes · 6.9k viewsRead it
Boz
@boztank · 2026-09-08
Very excited about the launch of @Muse today. I’ve been using it internally for months and I am hard pressed to think of any product that I’ve come to rely on more in such a short period of time. I have it linked to my email, calendar, and credit cards. I use it to help me plan travel, pack for trips, research and make purchases, and sort through all the communication I get from my kids’ school.
MuseThe company588 likes · 46k viewsRead it
signüll
@signulll · 2026-09-08
muse has been a genuinely impressive product to play with. it has all of the functionality of the imessage agents in flight today but also has all of the ingredients that will expose it to hundreds of millions including a massive friend graph through insta, increasingly rich context from email & other services which you connect, & more importantly stuff like fb marketplace. marketplace in particular is an incredible distribution wedge. millions of normal people could encounter muse simply cuz an agent helps them find something, negotiate the price, & arrange pickup. pretty good execution here from fb.
Muse1k likes · 70k viewsRead it
Loktar 🇺🇸
@loktar00 · 2026-09-08
Might have found a killer use for Meta Muse I use the HECK out of Marketplace (I know my GPU brethren also must). Sure Hermes can also do this, but if Muse is doing it it's got integrated access and no risk of being banned.
Muse56 likes · 38k viewsRead it
Danny Trinh
@dtrinh · 2026-09-08
Pro-tip: ask your Muse to organize your IG saves for you
Muse309 likes · 27k viewsRead it
JC Bahr-de Stefano
@jbahrdestefano · 2026-09-06
Nice going Instinct
551 likes · 1m viewsRead it
Jaya Gupta
@JayaGup10 · 2026-09-05
Wow meanwhile Instinct out here changing my passwords without asking
Instinct270 likes · 99k viewsRead it
Gregor Zunic
@gregpr07 · 2026-09-01
Here's how to build an Instinct clone 1. Host your persistent agent on @daytonaio 2. Make iMessage integration with @PhotonHQ 3. Give it browser use CLI from @browser_use 4. profit
Founder181 likes · 25k viewsRead it
Simon Taylor
@sytaylor · 2026-08-31
The consumer agent just started to feel real. I told Instinct: “You know that thing we did? Move it back 90 minutes.” It remembered the booking, the pub’s website, changed the reservation and sent me the confirmation. Most businesses are not ready.
156 likes · 229k viewsRead it
Rebecca Kaden
@rebeccakaden · 2026-08-31
Consumer agents have tipped from cool in theory to addictively essential to how I live and work. It's awesome. Now the 2 biggest things I crave (in order of importance): 1/ multiplayer functionality: Can my Instinct or Townie coordinate directly with my husband and friends' Instincts or Townies? Absolutely game changing (and wonder why more haven't prioritized multiplayer mode upfront.) 2/ figure out making phone calls. It's somewhat bizarre how much of the world still runs on phone calls but there are probably 3-5 times a week this comes up and I'm itchy for the agents just to integrate a way to do it.
193 likes · 49k viewsRead it
Olivia Moore
@omooretweets · 2026-08-30
Agents that actually work (esp. on mobile) are going to be a huge tailwind for voice AI I've massively increased attempts to voice dictate b/c of Instinct in the last week Instinct themselves could build the input, @WisprFlow will likely benefit too + we'll see what Apple does
83 likes · 18k viewsRead it