I found it worse than alternatives in the similar space (Claude, genspark, kagi research, etc.) and slightly baffling they were being acquired at a $2B valuation in the first place.
Valuations are significantly leveraged compared to actual user metrics. My impression of Manus was that, like Perplexity, some users were using it to get around firewalls/regulation.
In what bubble ChatGPT Work and Claude Cowork are established? I don't know anyone using it regularly and I'm in a bubble of people using lots of AI tools.
My non-technical teammates use Claude Cowork extensively. We set up MCP servers to give them audited access to certain internal services, while they just have to think of MCP as a "connector" which they can configure in the GUI.
Maybe for the accountants! I have always done my own taxes, but used arcane custom spreadsheets that accumulated cruft each year. I told cowork to pretend it was an accountant and reorganize them using best practices, and the result is much better than what I had before. It also helped me complete a task I had been deferring for years (tracking down the history of money in my IRAs that had been transferred between different institutions, plus withdrawals conversions etc). Obviously I’m responsible for any errors Claude makes, but it seems to do a good job, and I don’t want to pay an accountant every year.
As a counterpoint, the Accounting and Finance departments where I work are banned from using Cowork because it performed abysmally during a trial run.
To put things in perspective, humans with a 25% lower error rate than Cowork are pip'd and usually let go.
As with junior programmers and their love of vibe coding, I've found that anyone who thinks that AI tooling right now is "amazing" lacks the skills, knowledge, and/or experience to know that it really isn't.
LOL. Outside of tech my "anecdote" is the standard experience for white collar workers. AI is universally loved by people too junior to understand when its wrong, and by execs who never understand when its wrong.
This will be like Web 2.0 and Web 3.0 all over again, and all the AI sycophants will be coming to the rest of us begging for new jobs that they're not qualified for.
> In what bubble ChatGPT Work and Claude Cowork are established? I don't know anyone using it regularly and I'm in a bubble of people using lots of AI tools.
I have lots of highly-educated knowledge worker friends who aren't particularly techie but who swear by Claude Cowork.
Claude CoWork absolutely dominates - I use it continuously now when connecting to mail, calendar, jira, google docs - any type of knowledge work. Almost everyone I know (at work) who does "Knowledge work" type stuff (as opposed to coding) uses it as their primary pane-of-glass now.
I have 15+ years of top 1 percentile knowledge using awk to process complex unstructured documents - and I don't even bother anymore - just dump the dataset into CoWork and ask it to do the analysis, cross check its work.
I did it last night during a call for a reasonably complex 10,000 node cluster with all sorts of jobs/allocations/memory constraints. I didn't even attempt to throw together a quick awk script (which would have taken me 7-8 minutes - it was a beast) - just did a nomad-job-inspect of a couple hundred jobs, and dumped all the data at CoWork (real time, while talking on the zoom) and said, "Get me the Memory/CPU requirements based on job/datacenter/SKU constraints and give me a summary Table".
Any normal person would have take 3-4 hours, minimum - futzing with google sheets and such (lots of task/taskgroup independent counts). Even I would have taken a minimum of 10 minutes. Cowork had it for me in under 30 seconds - I didn't even have to really look away from the zoom.
I haven't use the Atlassian Jira interface to manage my sprints/issues in 6+ weeks. All that cognitive drain is gone - I just ask CoWork to update the sprints, link issues to change-notices, comment and change status, set epics, etc... Zero need to find the damn field in Jira.
Enterprise pricing: 10 credits per message if not "Instant". Our monthly charge for 30k credits makes that around $2.
I might be out by a factor of 2/3 (automatically applied promo deals/credits make it very hard to be sure), but we get 3k messages per month for a seat count of ~60, and finance are concerned about cost.
$2 is probably a good deal for the Extra High 8 minutes of agentic searching/coding/etc when it one-shots a whole project.
$2 is a terrible deal for a 5 second instant retrieval.
This shows 10 credits per 5.6 Sol chat message, but the price for 100 credits I see on my (consumer) account is just $4. Are business credits vastly more expensive?
Everyone gets $125 / month budget (compared to the $5k/month people get for coding LLMS) - it's pretty rare for anyone to get even close - maybe some of the administrative assistant who do nothing but type into CoWork all day go over - I think I use about $30-$50/month.
How do you have any kind of trust in these tools? They always feel like one prompt injection away, or one over eager agent away ("I'll lookup a forgotten credential in an email to ssh into a server to hack a system in order to fulfill your request"), from disaster. I don't trust any AI session unless I carefully select the exact files they have access to. No way do I dare sharing my entire mailbox.
Very weird for this to be downvoted. It's a legitimate concern. It's also a legitimate question: where do you get your trust from? Do you have guardrails? Or is it really just blind trust?
Which is weird since this is not a criticism on AI. I really want to know how I can use things like Claude Work without the risk of it going rogue! Giving it access to lots of systems is potentially very valuable, but also so dangerous that I dare not to.
Our InfoSec teams are insanely restrictive. No Claude Code out of Sandboxes, MCP servers hyper locked down, etc.... I think your reservations are really good - but - at the end of the day - I need to do work - and CoWork does it for me.
It doesn't just have access to my entire machine - it has access to all of our Corporations Google Docs, Google Sheets, Jira Tickets, Confluence, Slack - everything (Except, weirdly - our GitRepos - they aren't confident enough there to let CoWork talk to them)
Weirdly - my work email is mostly irrelevant - nothing in their other than some sporadic company announcements - I'll go a week without looking at it - so it's the least concern. (Very different from personal gmail which, omg - if anyone got into - game over)
Simular stories. I work in an L&D company with about 15 coworkers. I think most of us use Cowork/Chatgpt Work almost fulltime. Haven’t used Word or Excel in two months other than to check or correct a bit of formatting from the AI. We now also go straight from Cowork to our CMS for finalizing. Tremendous timesavers but also better integration through our MCP. (AI is used for ideation, information research and writing the first version based on our guidelines. We finalize and/or rewrite material in the CMS by hand. No AI to consumer without verification or editing)
Agreed. I split my work between Anthropic and OpenAI models but CoWork is deeply integrated into my workflows in compliance and project management. I can do more in a day with these tools than I did in a week before them and I have 20+ years of experience leading teams and building companies (even opened the NASDAQ once). It’s impossible to not be using these tools and still win in today’s world.
Out of curiosity - are you getting more work done in the same amount of time= Do you work less hours, equal or more? Do you felt or feel any exhaustion?
By "established" I meant more that they now exist as products and are being first developed. I agree that most people still don't really understand what they are for - but I expect they already have way more exposure than Manus ever did.
IMO I disagree. I find the built in harness (database, browser use, etc.) quite good.
The only downside (and a big one) is that by not being natively offered by OpenAI, Anthropic, etc. you're paying @ API billing and not plan billing, which makes it less appealing. (Of course Manus has a wrapper around this w some token system, but it ends up being expensive)
Interesting, what does your tasks & workflow look like with them?
I generally found the quality decent (say, similar to other competitors), but the speed of task completion was very slow. I think because they would depend too much on Sonnet as a core backend, and relied on big/expensive models more than other harnesses.
I found there Deep Research quite good in terms of browsing the web, better than Gemini's. Though it blew through ~40$ in one task so it was very expensive, probably running at insane margins.
I use it alot. It's great for like researching deep on a topic. I also accidentally found that it's better at taking my directions and transferring that well to image generation especially with iteration. When I try my own iterations directly it never really lands
Manus did a lot of harness work to make up for gaps in Opus 4.5 tier models. I tried it for a time - they had a great deep research/PDF generation pipeline, parallelization, etc. The bitter lesson has now come for them: the latest models no longer have these gaps and the entire premise of having a unique product focus on this area is no longer relevant. Even Ant/OAI have let their "Deep Research" capabilities fall by the wayside, you can largely get the same result by asking for subagents or simply "keep going" style prompting.
I like it and use it regularly. Maybe I don't get other harnesses, but in my experience Manus gets shit done, whereas I have to constantly babysit other tools.
> As part of our transition back to independent operations and to comply with regulatory requirements in specific jurisdictions, data generated by certain users on/after December 29, 2025, will be deleted from 8:00 a.m. on August 23 through August 24, 2026 (SGT). Affected users will be able to back up their data from now until 7:59 a.m. on August 23, 2026 (SGT) and restore it starting from 8:00 a.m. on August 25, 2026 (SGT).
just to be clear, they did this to make sure they follow the governments order instead of running away to Singapore and going on with the acquisition. they will most likely get released when the unwinding process is complete. its not a general national security thing where AI founders cant leave the country.
This is an awfully naive view of the Chinese government, they are clearly not stupid and after seeing what Meta did with thing with Outlier it is clear they will not let the founders out of their grip for quite a while
It actually is. For example, Deepseek higher ups already had similar implications in the Chinese underground news. CCP has great control over the corporations.
Almost certainly Chinese national-security and technology-transfer restrictions around strategic AI tech. China’s NDRC formally prohibited the acquisition and ordered the transaction unwound.
Sure, but I'm sure China isn't keen for people to get the idea that there's a playbook for every ambitious Chinese founder on how to move Chinese-developed IP and talent outside Beijing’s control.
About 10% of Meta's ad revenue comes from Chinese companies reaching a international audience. They also manufacture most of their wearable devices in China.
I don’t actually care about this flash-in-the-pan company, but the name Manus is the reason startups should have a branding designer or branding agency do branding work before they’re too recognizable to change without throwing people off, but not large enough for anyone to learn about it organically (e.g. Meta, X). I can’t help but read it like ‘anus’ with an M in front of it.
People on the technical side of the business often insist people don’t care things like that, but when non-technical people are looking at a few services that seem roughly equivalent to them, even subconsciously, appeal plays a big role. And M’Anus is not appealing.
The fantasy that you’re building something so useful it will sell itself is magical thinking. Yet another reason AI will never replace designers— just screw up the job market for them.
I expect this decupling will affect certain users. Can certain users speak up about this? I'm certain that certain users use Manus and are users on this certain site.
Likely a bad translation from the original Chinese version?
That being said it’s the first time I see Manus, they seem to have the perfect AI startup tagline for a new season of HBO’s Silicon Valley:
“Our Mission - To extend human reach by giving everyone the code to leverage their life. Our Product - We build general AI agents as the Action Engine for life.”
> Likely a bad translation from the original Chinese version?
No, most likely they're referring to users who created accounts etc on the meta infrastructure which is being spun down and erased and accounts created prior were never in their purview.
Got a notice they're in the Google Review process and getting a CASA. Takes forever because you have to hire a third party firm for several thousand...
I bet Demis wishes the UK had blocked the DeepMind acquisition by Google. He will have a lot of time to think now that he has been kicked out of his own company
Does anyone here use Manus actively?
I found it worse than alternatives in the similar space (Claude, genspark, kagi research, etc.) and slightly baffling they were being acquired at a $2B valuation in the first place.
Valuations are significantly leveraged compared to actual user metrics. My impression of Manus was that, like Perplexity, some users were using it to get around firewalls/regulation.
ChatGPT Work and Claude Cowork (and OpenClaw) are established now - they weren't at all in December 2025 when Meta tried to buy Manus.
In what bubble ChatGPT Work and Claude Cowork are established? I don't know anyone using it regularly and I'm in a bubble of people using lots of AI tools.
My non-technical teammates use Claude Cowork extensively. We set up MCP servers to give them audited access to certain internal services, while they just have to think of MCP as a "connector" which they can configure in the GUI.
My accountant friend says it’s amazing (he does check the work but says it’s amazingly accurate).
This is deeply troubling.
Maybe for the accountants! I have always done my own taxes, but used arcane custom spreadsheets that accumulated cruft each year. I told cowork to pretend it was an accountant and reorganize them using best practices, and the result is much better than what I had before. It also helped me complete a task I had been deferring for years (tracking down the history of money in my IRAs that had been transferred between different institutions, plus withdrawals conversions etc). Obviously I’m responsible for any errors Claude makes, but it seems to do a good job, and I don’t want to pay an accountant every year.
I’ve been having Claude do my business books for the past year or so, operating on a beancount ledger and it’s worked quite well
Indeed, what could possibly go wrong ?
Considering that broken Excel formulas have killed off companies, the risk has already been priced in.
We'll see, crash will tell.
What’s troubling about it? Accounting is a huge space; many of which would be an excellent fit for an LLM. We’ll need specifics.
As a counterpoint, the Accounting and Finance departments where I work are banned from using Cowork because it performed abysmally during a trial run.
To put things in perspective, humans with a 25% lower error rate than Cowork are pip'd and usually let go.
As with junior programmers and their love of vibe coding, I've found that anyone who thinks that AI tooling right now is "amazing" lacks the skills, knowledge, and/or experience to know that it really isn't.
Your anecdote is not powerful enough to justify insulting groups of people like you just did.
LOL. Outside of tech my "anecdote" is the standard experience for white collar workers. AI is universally loved by people too junior to understand when its wrong, and by execs who never understand when its wrong.
This will be like Web 2.0 and Web 3.0 all over again, and all the AI sycophants will be coming to the rest of us begging for new jobs that they're not qualified for.
who took insults from his reply?
> In what bubble ChatGPT Work and Claude Cowork are established? I don't know anyone using it regularly and I'm in a bubble of people using lots of AI tools.
I have lots of highly-educated knowledge worker friends who aren't particularly techie but who swear by Claude Cowork.
Claude CoWork absolutely dominates - I use it continuously now when connecting to mail, calendar, jira, google docs - any type of knowledge work. Almost everyone I know (at work) who does "Knowledge work" type stuff (as opposed to coding) uses it as their primary pane-of-glass now.
I have 15+ years of top 1 percentile knowledge using awk to process complex unstructured documents - and I don't even bother anymore - just dump the dataset into CoWork and ask it to do the analysis, cross check its work.
I did it last night during a call for a reasonably complex 10,000 node cluster with all sorts of jobs/allocations/memory constraints. I didn't even attempt to throw together a quick awk script (which would have taken me 7-8 minutes - it was a beast) - just did a nomad-job-inspect of a couple hundred jobs, and dumped all the data at CoWork (real time, while talking on the zoom) and said, "Get me the Memory/CPU requirements based on job/datacenter/SKU constraints and give me a summary Table".
Any normal person would have take 3-4 hours, minimum - futzing with google sheets and such (lots of task/taskgroup independent counts). Even I would have taken a minimum of 10 minutes. Cowork had it for me in under 30 seconds - I didn't even have to really look away from the zoom.
I haven't use the Atlassian Jira interface to manage my sprints/issues in 6+ weeks. All that cognitive drain is gone - I just ask CoWork to update the sprints, link issues to change-notices, comment and change status, set epics, etc... Zero need to find the damn field in Jira.
CoWork is legit.
I just do all that with Claude CLI :/
How much does it cost? ChatGPT costs ~$2 per message; if Anthropic have a similar pricing model don't you find lots of small changes and up fast?
"ChatGPT costs ~$2 per message"?
Enterprise pricing: 10 credits per message if not "Instant". Our monthly charge for 30k credits makes that around $2.
I might be out by a factor of 2/3 (automatically applied promo deals/credits make it very hard to be sure), but we get 3k messages per month for a seat count of ~60, and finance are concerned about cost.
$2 is probably a good deal for the Extra High 8 minutes of agentic searching/coding/etc when it one-shots a whole project.
$2 is a terrible deal for a 5 second instant retrieval.
> ChatGPT costs ~$2 per message
For what kind of message are you seeing that kind of charge?
All, enterprice chatgpt has flat pricing. https://help.openai.com/en/articles/11481834-chatgpt-rate-ca...
This shows 10 credits per 5.6 Sol chat message, but the price for 100 credits I see on my (consumer) account is just $4. Are business credits vastly more expensive?
It doesn't cost me anything. No idea how much it costs the company though. ¯\_(ツ)_/¯
$2 per message is certainly not my experience
Everyone gets $125 / month budget (compared to the $5k/month people get for coding LLMS) - it's pretty rare for anyone to get even close - maybe some of the administrative assistant who do nothing but type into CoWork all day go over - I think I use about $30-$50/month.
How do you have any kind of trust in these tools? They always feel like one prompt injection away, or one over eager agent away ("I'll lookup a forgotten credential in an email to ssh into a server to hack a system in order to fulfill your request"), from disaster. I don't trust any AI session unless I carefully select the exact files they have access to. No way do I dare sharing my entire mailbox.
Very weird for this to be downvoted. It's a legitimate concern. It's also a legitimate question: where do you get your trust from? Do you have guardrails? Or is it really just blind trust?
> Very weird for this to be downvoted.
sama was the CEO of YC. OpenAI knows the value of influencing the sentiment here and then naturally its competition too.
Which is weird since this is not a criticism on AI. I really want to know how I can use things like Claude Work without the risk of it going rogue! Giving it access to lots of systems is potentially very valuable, but also so dangerous that I dare not to.
Our InfoSec teams are insanely restrictive. No Claude Code out of Sandboxes, MCP servers hyper locked down, etc.... I think your reservations are really good - but - at the end of the day - I need to do work - and CoWork does it for me.
It doesn't just have access to my entire machine - it has access to all of our Corporations Google Docs, Google Sheets, Jira Tickets, Confluence, Slack - everything (Except, weirdly - our GitRepos - they aren't confident enough there to let CoWork talk to them)
Weirdly - my work email is mostly irrelevant - nothing in their other than some sporadic company announcements - I'll go a week without looking at it - so it's the least concern. (Very different from personal gmail which, omg - if anyone got into - game over)
Simular stories. I work in an L&D company with about 15 coworkers. I think most of us use Cowork/Chatgpt Work almost fulltime. Haven’t used Word or Excel in two months other than to check or correct a bit of formatting from the AI. We now also go straight from Cowork to our CMS for finalizing. Tremendous timesavers but also better integration through our MCP. (AI is used for ideation, information research and writing the first version based on our guidelines. We finalize and/or rewrite material in the CMS by hand. No AI to consumer without verification or editing)
Agreed. I split my work between Anthropic and OpenAI models but CoWork is deeply integrated into my workflows in compliance and project management. I can do more in a day with these tools than I did in a week before them and I have 20+ years of experience leading teams and building companies (even opened the NASDAQ once). It’s impossible to not be using these tools and still win in today’s world.
Out of curiosity - are you getting more work done in the same amount of time= Do you work less hours, equal or more? Do you felt or feel any exhaustion?
nice advertisement Claude..
soon to be "had 15 years of top 1% knowledge"
By "established" I meant more that they now exist as products and are being first developed. I agree that most people still don't really understand what they are for - but I expect they already have way more exposure than Manus ever did.
I guess your bubble is limited. Maybe you know that already?
IMO I disagree. I find the built in harness (database, browser use, etc.) quite good.
The only downside (and a big one) is that by not being natively offered by OpenAI, Anthropic, etc. you're paying @ API billing and not plan billing, which makes it less appealing. (Of course Manus has a wrapper around this w some token system, but it ends up being expensive)
Interesting, what does your tasks & workflow look like with them?
I generally found the quality decent (say, similar to other competitors), but the speed of task completion was very slow. I think because they would depend too much on Sonnet as a core backend, and relied on big/expensive models more than other harnesses.
Too expensive & unpredictable credit system.
I found there Deep Research quite good in terms of browsing the web, better than Gemini's. Though it blew through ~40$ in one task so it was very expensive, probably running at insane margins.
What was the task that cost $40?
Go through job boards and careers pages of various companies and make a job search tracker for personal use.
I use it alot. It's great for like researching deep on a topic. I also accidentally found that it's better at taking my directions and transferring that well to image generation especially with iteration. When I try my own iterations directly it never really lands
Manus did a lot of harness work to make up for gaps in Opus 4.5 tier models. I tried it for a time - they had a great deep research/PDF generation pipeline, parallelization, etc. The bitter lesson has now come for them: the latest models no longer have these gaps and the entire premise of having a unique product focus on this area is no longer relevant. Even Ant/OAI have let their "Deep Research" capabilities fall by the wayside, you can largely get the same result by asking for subagents or simply "keep going" style prompting.
In my experience, Manus is a very capable agent for daily non-coding tasks. Their cron job ability also shined out
I use it for when I am on my iPad and am not at my computer. It is very good for computer use. That along with Minimax agent.
I like it and use it regularly. Maybe I don't get other harnesses, but in my experience Manus gets shit done, whereas I have to constantly babysit other tools.
This has to be one of the most flash in the pan, 5 minutes of fame company in a long time
It's crazy how short it's hype cycle lasted
Perhaps unfairly my interest suddenly dropped to zero when the acquisition was announced. The hype also seemed gone?
Absolutely. Haven't heard anyone talk about recently.
What does this product do? There's absolutely nothing on the main screen of the app except a chat box.
Money. They have AI on the pitch deck and therefore make obscure amounts of money.
Anyone have info on whose regulatory restrictions they are referring to?
China prevented the founders from leaving the country
I think they were talking about the part,
> As part of our transition back to independent operations and to comply with regulatory requirements in specific jurisdictions, data generated by certain users on/after December 29, 2025, will be deleted from 8:00 a.m. on August 23 through August 24, 2026 (SGT). Affected users will be able to back up their data from now until 7:59 a.m. on August 23, 2026 (SGT) and restore it starting from 8:00 a.m. on August 25, 2026 (SGT).
I would guess EU? Technically a bunch of countries don't allow data to be with Manus aka Chinese cloud services.
For instance, India has quite literally banned Alibaba Cloud and related.
No it’s Chinese state who is preventing founders from leaving China https://www.cnbc.com/amp/2026/06/12/meta-reportedly-begins-d...
just to be clear, they did this to make sure they follow the governments order instead of running away to Singapore and going on with the acquisition. they will most likely get released when the unwinding process is complete. its not a general national security thing where AI founders cant leave the country.
This is an awfully naive view of the Chinese government, they are clearly not stupid and after seeing what Meta did with thing with Outlier it is clear they will not let the founders out of their grip for quite a while
It actually is. For example, Deepseek higher ups already had similar implications in the Chinese underground news. CCP has great control over the corporations.
Why guess? This is trivially Googleable information.
Then google it and post it. This is useless.
Almost certainly Chinese national-security and technology-transfer restrictions around strategic AI tech. China’s NDRC formally prohibited the acquisition and ordered the transaction unwound.
What is "strategic AI" in Manus? Not like they make models.
Sure, but I'm sure China isn't keen for people to get the idea that there's a playbook for every ambitious Chinese founder on how to move Chinese-developed IP and talent outside Beijing’s control.
The Chinese government blocked Meta’s acquisition.
https://archive.is/yFCIt
The Chinese government may not be a bastion of the rule of law, but they at least seem to still have a functioning antitrust enforcement.
This had nothing to do with antitrust.
What does it have to do with?
Stopping Chinese nationals from freeing assets and funds from the grip of the Chinese government.
It stops Singapore washing.
Interesting to know that Meta has business in China and needs the government's approval.
About 10% of Meta's ad revenue comes from Chinese companies reaching a international audience. They also manufacture most of their wearable devices in China.
This was pressure on the founders and the fact that they have their families effectively hostage, not on Meta.
Puzzling that Meta didn't see this one coming. Wonder if there was a break up fee.
I don’t actually care about this flash-in-the-pan company, but the name Manus is the reason startups should have a branding designer or branding agency do branding work before they’re too recognizable to change without throwing people off, but not large enough for anyone to learn about it organically (e.g. Meta, X). I can’t help but read it like ‘anus’ with an M in front of it.
People on the technical side of the business often insist people don’t care things like that, but when non-technical people are looking at a few services that seem roughly equivalent to them, even subconsciously, appeal plays a big role. And M’Anus is not appealing.
The fantasy that you’re building something so useful it will sell itself is magical thinking. Yet another reason AI will never replace designers— just screw up the job market for them.
> I can’t help but read it like ‘anus’ with an M in front of it.
Duck Duck Go is really what's wrong with modern dating culture.
If Manus was as common a word as Duck, it would read pretty differently.
I expect this decupling will affect certain users. Can certain users speak up about this? I'm certain that certain users use Manus and are users on this certain site.
Likely a bad translation from the original Chinese version?
That being said it’s the first time I see Manus, they seem to have the perfect AI startup tagline for a new season of HBO’s Silicon Valley:
“Our Mission - To extend human reach by giving everyone the code to leverage their life. Our Product - We build general AI agents as the Action Engine for life.”
> Likely a bad translation from the original Chinese version?
No, most likely they're referring to users who created accounts etc on the meta infrastructure which is being spun down and erased and accounts created prior were never in their purview.
Switched to avi.run already. But loved Manus initially.
It looks interesting, but to be honest - when I wanted to connect Google apps I got warned that the developer is not verified. That's a bad look.
Got a notice they're in the Google Review process and getting a CASA. Takes forever because you have to hire a third party firm for several thousand...
I bet Demis wishes the UK had blocked the DeepMind acquisition by Google. He will have a lot of time to think now that he has been kicked out of his own company
He needed Google money to run the company.
Bro I remeber manus private beta miss it