What this covers
Nate B Jones reviews Comet, Perplexity's new $200-per-month AI agent, after months of testing competing tools. The core argument is counterintuitive: Comet succeeds not because of superior AI capability, but because of superior interface design—specifically, letting the assistant operate invisibly in the browser sidebar rather than forcing users to supervise its every step. Jones tested Comet against eight different workflows spanning flight reservations, restaurant searches, LinkedIn triage, Gmail drafting, and calendar management, demonstrating how it completes end-to-end tasks like meeting rescheduling without requiring the user to open multiple applications or watch the agent work.
The review situates Comet against two categories of failure. Competing task builders like Zapier and n8n require users to manually define and configure agent behavior upfront, limiting them to specific workflows rather than general-purpose assistance. Operator, OpenAI's entry, has more powerful AI but wraps it in a slow, cramped toy-sized browser that kills adoption. Jones argues the real strategic prize is owning the browser itself—the operating system layer where people now spend their working lives—and that Comet's native integration with Gmail, calendar, and LinkedIn gives Perplexity that foothold.
On valuation, Jones proposes measuring Comet's worth by aggregating the five- to fifteen-minute slices it saves throughout the day; at a $35-per-hour rate, ten hours saved monthly yields close to 2x return on the $200 subscription. He recommends prospective users commit to one month of disciplined testing in their own environment before deciding whether the accumulation of micro-savings justifies the cost. Jones also notes that Comet arrived three or four months ahead of when he expected production-ready general-purpose agents to exist.
The reviewer argues that Comet by Perplexity is the first genuinely useful general-purpose AI agent because its winning differentiator is UI design—letting the assistant disappear and do the work—rather than raw AI capability.
- Competing agents (Operator, Zapier, n8n) fail or limit users because they burden them with building/supervising the agent or have poor UI
- Comet integrates natively with browser, Gmail, calendar, and LinkedIn so tasks complete in the sidebar without context switching
- Value should be measured by aggregated time savings, justifying a new $200/month software valuation paradigm
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
All major AI players are converging toward owning the browser because we live so much on the web that becoming the dominant browser of choice effectively makes you the operating system for AI—the strategic prize Perplexity is pursuing with Comet.
“the key if you're perplexity is that you are making an OS, right? This is where all of these AI players are going.”
Comet is a true general-purpose agent assistant, with the closest prior comparisons being OpenAI's Operator and Gemini's Project Mariner, but Operator's vision (ChatGPT decisioning across the web) was doomed by poor execution and UI—a slow, awkward toy-sized browser inside a chat window.
“Comet by Perplexity is really a generalurpose agent assistant. The closest comparison we've had previously has been Operator from OpenAI and Project Mariner”
Whether Comet is worth $200/month requires disciplined measurement: adding up the 5–15 minute increments it saves throughout the day; if it saves 10–15 hours a month it is clearly worth it, and at a $35/hour valuation 10 hours saved yields close to 2x ROI monthly—representing a new valuation paradigm for software driven by baked-in cognitive intelligence.
“whether it's worth $200 a month essentially requires you to add up those 5, 8, 10, 15 minute increments that it's going to be saving you and be disciplined about it, measure the value of your time”
The widespread design assumption that users must see what the agent is doing is a legacy of treating agents as untrustworthy; once an agent can hook into data and just works, close supervision becomes unnecessary and undesirable—users want it to be fast, trustworthy, and dependable.
“the idea that we have to see the agent is probably a legacy of the idea that the agent is untrustworthy”
When prompted with the thesis that TikTok Shop's slowdown affects social commerce broadly and may impact Amazon's outlook, Comet's analysis concluded Amazon is pursuing an agnostic social commerce approach—explicitly responding to TikTok with integrations across Instagram, Snap, Pinterest, and YouTube Shorts—and is hedging risk and investing further rather than ignoring the slowdown.
“Amazon's produ pursuing an agnostic social commerce approach explicitly response to Tik Tok. uh similar integrations with Instagram, Snap, Pinterest, and YouTube Shorts.”
Comet works as an AI agent not because of superior AI but because of superior user interface—the team understood that the user's goal is to let the assistant do the work and disappear, rather than to watch or micromanage everything the assistant does.
“Comet is that AI agent. And the reason why is not AI, it's UI, it's user interface.”
Before recommending Comet, the reviewer subjected it to eight different workflow tests across diverse tasks (flight reservations, restaurants, LinkedIn, Gmail, calendar), recommending others sign up for one month and rigorously test it in their own environment, honestly measuring time savings before committing to the investment.
“I put Comet through eight different workflow tests this morning just to see how good it is before making this video. I was not easily impressed.”
Zapier and n8n both lean users into building an agent by requiring them to define exactly what it should do; n8n is rising in popularity because it does this better, but neither is a general-purpose agent—they are useful only for specific agentic tasks like extracting text from a document into a spreadsheet.
“Both of them lean you into building an agent. You have to define exactly what you want it to do.”
Because Comet operates natively within the browser, it can help users make better decisions in a way screenshotting cannot—there is something about native assistant-browser collaboration that cannot be replicated cleanly elsewhere, representing additional hard-to-measure value Perplexity is positioned to capture.
“because it can look at the browser natively, it can help you make better decisions in a way that like screenshotting doesn't really do”
Many current AI agents bury the user in the cost of building or directly controlling the assistant, forcing heavy upfront investment in defining and configuring behavior, which undermines their value as general-purpose tools.
“so many agents right now bury you in the cost of building the assistant or controlling it directly”
Comet's browser-takeover UI—glowing the main browser blue while it acts, then removing the glow—is an intuitive signal that the agent is driving while preserving the user's ability to stop it at any time.
“it actually goes to your main browser and glows it up blue like you're like looking through a Star Trek portal or something and then it like does something quickly in the browser and then like it takes the blue away”
Comet completed a real meeting-rescheduling workflow end-to-end from the sidebar: it found the meeting, suggested a better time fitting the user's preference for clean working blocks, drafted a calendar change, then proactively recommended sending a notification email which it also drafted—all without the user opening Gmail or Google Calendar.
“It actually gave me a recommendation for a better time within that block that would fit well with my goal of like clean working blocks. And then it drafted up um a calendar change.”
Comet can run multiple tasks simultaneously across integrated services—LinkedIn, calendar, Gmail—and a running task continues even when the user navigates elsewhere in the address bar.
“It plugs into LinkedIn. It plugs into your calendar. It plugs into your Gmail. You can do multiple things at once. This is going to keep running even if I change where I'm going up here in the address bar”
Comet feels three or four months ahead of schedule—the reviewer expected general-purpose agents were coming but did not expect one this fast, efficient, effective, and connected across flight reservations, restaurants, LinkedIn, Gmail, and calendar this soon.
“This fast and this connected feels three or four months early.”
Comet's response on Amazon's social commerce strategy was more coherent, actionable, and on-point than typical AI output, which tends to be bland.
“I really appreciate this is actually a more coherent take than I typically see from AI at this point. It's it's not just bland”
The reviewer suspects Operator misreports its task time, claiming an eight-minute job actually took around 20 minutes of clock time, possibly because it starts the clock late.
“I've had instances where operator says it took eight minutes and it was like 20 minutes of clock time.”