This is a wild statement that does not seem to be supported by any actual data.
What does it mean? Does clicking on a link counts as labor.
What they will have done is asked a human who should be knowledgeable approximately how much time they spend on the activity (e.g.: how long do you normally spend copy and pasting? How long do you spend looking for the files you need? etc.). When you ask someone whose job it is, they tend to overestimate, and on top of that, you break down the questions as much as possible, so these small overestimations compound without it being obvious to the person that they're making a mistake. It's easy enough to spot that you've said 'four hours' but you know a task doesn't to take you a full morning.
Once you've got all these answers, you ask how often the person has to do it.
Then you ask someone in HR for an average salary you can use. Now that AI is doing that work, you multiply the number of hours saved by the average salary, and report that as your savings.
Something, as usual, stinks about these numbers. $1.6M in saved labour costs, and 3.25 years of work in 4 weeks are basically two ways of saying the same thing (labour costs vs improved productivity).
So let's say 3.25 years of work is around 160 weeks, minus the four weeks they actually took, so 156 weeks of productivity savings. Assume 40 hour weeks, that's 6,240 hours 'saved'. Which works out at around $250/hour, which... well. You decide if that's plausible.
I think we might be seeing what happens when people are being paid too much to spend all day emailing each other and jockeying excel/gantt charts/org charts. Yeah for some definition of "work" I guarantee that a LLM could perform 3.25 years worth in four weeks.
We should definitely seek ways to turn things over to AI. Then turn the AI off, and figure out what actually needed to be done.
> people are being paid too much to spend all day emailing each other
Hmm, this does not sound exactly right. Also, does anybody seriously think that communication is not work, or is not important? A number of really impactful things started from people emailing each other. (Hell, Linux kernel development is still much about people emailing patches each other.)
The lingering question is if the intermediate LLM translation steps will actually make our communication more efficient - or just amplify the already inefficient parts.
For example, one task takes a document with data, charts, and metrics, and Perplexity Computer was tasked with creating a 10-page slide deck for a presentation. Prior to AI, that took human capital and labor costs.
I can't say whether the $1.6M in labor costs is legit or not, but these tools are not just clicking links in 2026.
send me the data and ill ask my own AI to do it in my favorite silly voice.
I want to know pre-"personal computer by perplexity"
I think their numbers of $1.6M and 3.25 years is still probably a massive overestimate, but the order of magnitude seems plausible.
The typical market research , Google analyze , put into spreadsheet is almost gone job. Imagine how many people were doing that as major part of their work
Other example is "finding" employees. What's the purpose of a human in the middle? The implied result is contact list for potential employees. Will they write them invitation for the interview by hand?
I'm really confused why is this presented like this. It has this surreal dreamlike quality. Something is done. Emails sent, contacts acquired, profits unlocked, synergy achieved etc.
AI marketers are leaning into use cases where and individually can unilaterally choose to do their work a different way. This is much easier to explain and distribute!
All that said, the result is definitely funny - seeing work done like this makes you realize how artificial the tasks that make up modern work are.
You are missing the point of board briefings. The CEO serves as a critical filter of information, deciding what to tell the board and how to frame it. If you take the CEO out of it, you're giving the board full access to the company's state. There's enough going on day-to-day that each member can tell themselves the story they want to believe. The CEO is there to advocate for the company and present a unified front, you can't take them out of the equation.
What does this mean? The computer isn't alive. It's physically located on my person? Phones and watches have already cracked this.
If I say "Bob lives with me", that just mean that they generally share a residence with me. Desktop PCs already do that.
I just don't understand what's even intended by this.
But they want you to think of it as alive. They're anthropomorphizing it.
I might be misinterpreting, but according to the landing page, this is the intention:
> Personal Computer gives Perplexity Computer and the Comet Assistant always-on, local access to your machine's files, apps, and sessions through a continuously running compact desktop.
> It's a persistent digital proxy of you. Controllable from any device, anywhere.
That being said, the grandeur and bombastic language also seems fitting for something less sinister, like an even worse version of MS Recall maybe? Combined with, let's say... agents!
That's it! You Personal Computer is your agent and not only may act on your behalf, it also communicates your preferences and intentions.
Futuristic, right?
Would a real person risk their reputation like that?
--
With regard to the attempted redefinition of a commonly used term, I'm reminded of Gretchen, from the Mean Girls, trying to redefine "Fetch!"[1]
It's just not going to happen.
https://www.fastcompany.com/91497841/meta-superintelligence-...
… particularly with acts that have legal implications like … well, almost everything, but particularly communication with investors or board members.
If people can get slides or summaries by pushing a button, they don't need others to push the button for them.
The slide deck won't be viewed by a human. It'll be read by the human's pet LLM and then summarised into 3 bullet points.
This seems like a good device for working on a single project of well-defined scope. You can give it whatever context it needs by including the relevant data, and no one's asking you to log into your Apple or Google account to go through your private photos. Keep separate things separate.
Maybe in a couple of iterations, you'd be able to trust the AI to straight up drive your computer with access to all important parts of your digital life most of the time and only occasionally have to manually stop it from wiring all your savings and 401k to a struggling Nigerian prince.
>Personal Computer runs on a dedicated Mac mini that can run 24/7, connected to your local apps and Perplexity’s secure servers.
Choose Perplexity Computer if you: want a managed, safer, minimal‑setup agent for research, content, presentations, and business workflows, and you’re fine paying a subscription for a polished cloud experience.
Choose OpenClaw if you: are technical, want local code execution and device automation, prefer full control over models/tools, and are willing to own the security and troubleshooting burden.
One thing I noticed is that whatever harness PPLX wraps around the models, the output is noticeably lower quality in aggregate. I assume some kind of token compression being used before passing your query to a given model but to my knowledge that's never been proven or confirmed?
Anyways, I get the most value out of coding and PPLX has seemingly pivoted away from that. Probably a good play to not try and compete directly with Claude Code/Codex and find a better niche, but I am not sure who or what their market is. Lovely design, however.
- Perplexity: This one has been promoted on (insert general audience media skewing toward the older set) enough to be a household name still.
- ChatGPT: General people in some demographics (see immediately above) are averse to this, on account of negative publicity its parent company has received. (Still very strong popularity and positive sentiment in some demographics, though)
- Claude: Some semi-literates have glommed onto this one, possibly as a result of its more recent success among the developer set.
- Grok: People can be either for or against, based on how they feel about its owning company and its ownership; no more need be said
- Gemini: Again, if you are in the universe of its owning company (or decidedly not), the draw (or repulsion) can be strong here.
For general LLM use, the above are all about the same. To be clear, this is just me shooting from the hip for how each offering might be viewed. IMO, it's not a bad idea to submit the same input to each and see how they compare, if one is so inclined.
So a more polished OpenClaw that integrates with Perplexity?
In general interesting, if it's not just limited to Mac Minis. Would love to put this on my VPS that's currently running OpenClaw
Also this "system" just seems vulnerable af.
The broader trend is pulling back a bit on “minimalism,” right? I think we hit peak (or valley?) minimalism already so I guess there’s only one way to go.
However, in my opinion this specific typeface and aesthetic is been taken up by AI companies to harken back to the likes of the 1984 Macintosh ads and such...in an attempt to try and convey that "$(AI_PRODUCT) is just as revolutionary as the first desktop PCs".
Build everything, do anything, give AI all your data and thoughts and system access and it will give you the world!
I'm not surprised our own "roaring" 20s is seeing this shift.
No, it doesn't, because it's not alive.
I thought of zombo.com the other day and booted it up. There is maybe no other website that continues to bring me as much joy as zombocom
...because this thing will go rogue faster than you can blink.
I swear, it's like nobody at the company even reads the slop they're generating or thinks about it for any amount of time. In what world is advertising a kill switch as one of its essential features a positive? It's basically admitting from the start that this is unreliable.
There's a sense of "early bitcoin" around clawbot and other agent frameworks. I think if you wait for another 2 years for it to mature, you'll have missed out as if you waited ten years after bitcoin began.
They're insecure and janky, sure, but on the other hand you've got millions of dollars of compute and tens of thousands of very motivated developers working on making them secure, reliable, and competent. There's something magical about AI that actually gets real work done while you're doing other things, and that's what Perplexity is probably hoping to sell.
Just need a reliable local model, though - AirLLM, other hacks allow you to run bigger models more slowly, so you can build out a completely API-free scheme to run pretty capable agents even without big GPUs.
Could be a Moravec's paradox thing - all these people are thinking that the solution looks enticingly within reach, but it might be an absolutely horribly complicated quagmire with no easy solution short of AGI. I'd bet on clawbots and agents being very secure and great to work with in the very near term, though.
It’s been 2 months.
https://www.perplexity.ai/hub/blog/everything-is-computer
They designed a program (copied OpenClaw) and called it a computer
They may not come after all the niche companies, but they definitely come after the most successful markets, especially those with low effort moats.
Same goes for relying on the Apple/Google app stores (ex - Apple literally got slapped in court for copying successful apps and then pushing their offering to the top of their stores... talk about wildly abusive behavior).
I may still choose to use AWS/GCP/Azure while trying to find product-market fit as an immature startup, but I'd look real, REAL hard at ditching them as soon as possible afterwards.
Unless you have particularly bursty workloads, they aren't even a good cost saving measure anymore.
* The first is where there's you and your computer, and you're doing pre-AI work. You hit some hotkey and pass off some task to AI.
* The second - and where I think we should probably be going, is there's you and you interact with the an agent. You aren't handing the Q4 report off to the agent, the agent is bringing the Q4 report to you.
I think the first scenario is trying to pry agentic work into legacy workflows. It will be more powerful when we simply go straight to the second, where orchestration and interaction with your agents is the interface.
It's difficult to understand what this is because its name is "Personal Computer", and it seems like their definition of Personal Computer is very different from everyone else's.
Also it's funny that it shows making a revenue report with their brand template. AI can replace HR jobs but they still have to make reports for noble executives? They are basically saying "We won't replace CEOs/executives".
If Perplexity ships this successfully, I suspect it could push the whole space forward. Once a big company normalizes the idea of an always-on agent with OS-level access, we’ll probably see a lot more companies building similar “AI computer” layers on top of existing systems.
Seriously though, Perplexity, like most of the AI wrapper companies, seems unable to innovate much beyond the query-response chat paradigm. I don't understand why VCs continue to fund these ai-slop companies. I see a new company's advertisements on the NY subway every week, and they're all the same: Anthropic/Google/OpenAI resellers who are selling some UI wrapper (or at best a bespoke model worse than the flagships) on top of pretty basic prompt engineering or tools.
This is what happens when we invert the product-paradigm: we're not solving problems with technology, we're taking technology and applying it to problems.
I use AI every day, so I'm hardly a luddite, but this bubble is so ridiculous at this point. This perplexity product, more than any other so far, feels so representative of peak craze.
Back when work was work and we had festivals.
If you went to the right school, you can be a Series A company with nothing more than an OPENAI_API_KEY. Most of the young and inexperienced founders mentally retire at this point and start their family planning.
There were a few cities like "Austin", but now I guess it's just SF and NY again (I'm in SF).
The students get what they paid for.
The school gets their metrics.
The VCs get returns, from the increased revenue their portfolios just paid for by investing in their kid's startup.
And the wheel in the sky keeps on turnin!
> Watch video on YouTube > Error 153 > Video player configuration error
On Firefox/Linux or Safari/iOS.
How is it that this kind of organisation can't properly embed a video player or make a working landing page?
I would be willing to try this new product of theirs, but definitely on a secondary computer (i.e. not main system).
Do I have to sign up to install their version of an OS/openclaw?
>Depends on our SaaS
Pick one.
Given the inherent unpredictability of LLMs, I'm not convinced that an openclaw-like system but with more security features bolted on top is really a positive in the sense that the false sense of absolute security probably outweighs whatever actual security has been added.
It is easier to understand that openclaw is definitely insecure.
I don't think I'm cut out for the modern world
Basing this concept on what we have today with LLMs is a call for chaos, unreliability and slop communication; at best.
Uh... how about..., no...? What?
> Controllable from any device, anywhere
...
> There is a kill switch
Oh great, sounds like you're confident this is safe then!
Someone wake me up, please
Hasn't this been true since, like, the late 70s?
“The Browser Company”
What’s next “Car, by Ford”
Brand-wise it’s like those cheap dishes at Ross that have the word of what it’s for, like a tea cup that just says “Sip” or a jar with the word “Tea”.
I hate it!
AI brands should cut to the chase, something edgy like “Villain” or “Blackbox”
