Grammarly acquires Superhuman | Hacker News Reader

Grammarly acquires Superhuman

reuters.com

188 pointsthm133 comments

Superhuman user and former Grammarly user here.

I'm a big fan of Superhuman as an email client and happily pay the premium price for it. I really hope they don't change what makes it great.

I used to love Grammarly until they essentially ruined their product - much like Dropbox did. They took an app that worked perfectly and deprecated it, replacing it with an invasive keyboard replacement that was supposed to work everywhere but performed poorly across most programs and included functionality I wasn’t interested in that is kept nagging me to use. When I complained about the issues, instead of addressing my concerns, they sent form letter responses about their commitment to privacy rather than fixing their intrusive software.

This reminds me exactly of Dropbox's transformation from simple, reliable file storage into bloated software that cluttered my computer with pop-ups and background processes. When users complained, their team never seemed to understand why we were frustrated. Then they started acquiring other services I eventually cancelled as they tried to integrate them into their core service. I eventually moved to iCloud and never looked back.

I hope Superhuman keeps their current excellent email client that I gladly pay for, rather than replacing it with some "next generation" product that nobody asked for and that would likely be inferior to what we have now. I genuinely love Superhuman as it is.

What do you like about Superhuman that Spark Mail doesn't have?

So my work is client based and it really makes sense to divide almost all the email into client folders. Superhuman lets me quickly move stuff from my inbox into the client’s folder. It’s just very very quick.

I actually used spark mail for a while, because it would export to my crm, but eventually I wrote a script that converts all my emails to pdf to drop into my crm.

> Then they started acquiring other services I eventually cancelled as they tried to integrate them into their core service

I was struck by a snippet I read recently - I can't remember the organization being discussed - that I'll paraphrase as "The company is the product, and it's being sold to shareholders", and currently, it is very fashionable to sell the story of unceasing growth, so companies will do anything, including turning away existing customers, just to have a shot at growth so the share price can keep rising.

> unceasing growth

As someone with a finance background and job, this deeply annoys and irritates me. Even made me quit a market finance path for something more realistic. As the saying goes, "you can't have infinite growth on a finite planet". I don't know why we need to explain something so obvious to people who (1) have the power to make and (2) are making decisions with large scale impact, every minute. Unless there's something (pure evil, maybe?) I haven't uncovered yet.

So much software keeps repeating this pattern: they gain their initial success by finding something users want to do, and then building software to enable users to do it. Users love it and are happy!

Then, after they’ve enjoyed some amount of success, they try to flip the script: now they start with what they want users to do, and build software that encourages, annoys, cajoles, or insists that users do that thing, even if they don’t want it!

So much software falls down this pattern and just endlessly begs the user to change their behavior, rather than simply addressing the users’ actual needs.

I think the phrase you are looking for is called enshitification.

Coda has some great folks leading it, so let's hope that now they are really running grammarly, they can hold out against this.

The answer will depend on how well they do at taking care of the $1B line of credit used for the purchase.

Grammarly has an existential crisis. It can be replaced it with free versions of the top models and they are much better (and I can control the UI anyway I want). In fact, many of these “web 2.0” business models are a few more updates away from getting replaced.

Grammarly can cut down their costs if they use those models themselves. The current LLM advancements aren't disruptive but incremental. What's the hurdle you see they can't rely on their existing distribution and expand from there?

That’s true but I can replace Grammarly in a few hours with Claude. 99% of the functionality. Then, I can spend a couple more days to add stuff that they can’t add due to copyright.

Grammarly has a function that replaces phrases that are typical for LLM output. Good luck coding that in a few hours.

Last I read, they run it off of MSFT/OpenAI.

I'm sure they haven't created their own model. Probably fine-tuned it + constantly updating a database of the most occurring phrases from all LLMs.

Why would using a more-expensive technology cut their costs?

It's the dot-com strategy. They'll lose money on every sale but make it up in volume.

not to mention languagetool

yet they recently received a billion dollars in financing this year?!

Yes, does not seem like such a crisis to me.

maybe not that crisis, but perhaps a crisis

I'm paying for most AI models top tiers and Grammarly. Grammarly is a phenomenal tool. It's not that LLMs can't do it. Well, they can't, but the more important thing is Grammarly's UI.

What do you like about their UI and what are your main use cases?

I can see room for improvement about the UI, but I like that it works everywhere and I use it to, obviously fix grammar and mistakes, to add/change our company's specific vernacular (finance), and to change common LLM sentence forms so that the readers don't notice that content was made by AI. Ya know, all these em dashes, "testaments to", "this isn't just X, it's Y", etc.

My favorite function of the UI is that, on any website, pressing TAB just fixes/improves what I just wrote before.

Agree with this. I love that Grammarly is integrated into everything, and I don't have to switch to an AI chat for minor text edits.

Exactly. Also, it's incredibly affordable, especially when you buy it on Black Friday.

This is such a programmers take. Bottled water is in a crisis, it can be easily replaced with tap water and a reusable container! Yet it’s a $47 billion dollar market in the US.

Grammarly’s value is not in having a replaceable product, it’s in the network, distribution, customer acquisition channels and integrations with tools. Like bottled water, it’s about being in the customers face at the right place and the right time.

You’re on the money and that’s exactly why I think Grammarly will struggle. OpenAI/Gemini/Claude will get embedded further. Gemini is already on gmail. Getting OpenAI and Claude incorporated is trivial. Guess what? Once Apple figures out what they’re doing with AI (which I hope is to buy Anthropic) they will take whatever is out there and incorporate them into iOS/MacOS just as they’ve done for so many third-party app ideas in the past.

I've been looking to replace Superhuman recently. None of their AI or Team features matter to me. I just wanted what they originally set out to build -- a super fast, keyboard driven, desktop email client. There are daily paper cut bugs and search issues that have persisted for many years, and I'm not going to stick around through this transition which will surely make the product worse.

What do folks like for desktop email that's keyboard driven? At this point I almost want to go back to Pine ;)

Mailmate on macOS is good

Mimestream on Mac implements GMail's keyboard shortcuts so there was no learning curve for me and I'm able to enjoy a desktop app experience.

This does seem pretty seamless to swap to, thanks!

Now to figure out iOS ;)

chiming in with thanks, i was looking to leave Superhuman also.

Is there a reason you prefer desktop apps over using email in the browser?

Mimestream user here: I can command-tab to an application.

Same here, I recently got frustrated with every email client, tried everything. Few days ago I finally decided to vibe code my own email client with Claude Code and I got a basic version running in a single day. Can’t wait to build it exactly the way I want, with programmable rules/filters and AI drafts for specific types of emails I get, conversion to plain text or dark mode for readability, contextual information in sidebars pulled from APIs (such as email history and customer support / CRM info), and one click actions in other apps.

I use (neo)mutt

In the same boat. Tried Notion Mail and it wasn't quite at Superhuman's level, though I do anticipate they'll improve it.

We're building what you want:

https://marcoapp.io

are you unifying the inbox across multiple mail accounts? (like apple Mail.app does)

Yes, that's one of our core USPs. It's possible to view one email account at a time, all at once, or any user-generated combination you might desire.

I’m using Shortwave and it’s nice! They’re also leaning into AI unfortunately but the core email experience is pretty good with nicer notification control and bundling features than Superhuman had.

After a day of trialing many apps, briefly settling on Gmail (iOS, and Unite to make a desktop app), I'm pretty happy with Spark for MacOS and iOS.

Simplehuman for Gmail is what I use. It's a light weight browser extension that makes Gmail work like Superhuman with the same keyboard shortcuts and natural language snooze. https://simplehuman.email

For fast, combined Google accounts with natural language prioritization, we're building https://wyntk.ai

> The company claims its users send and respond to 72% more emails per hour, and the percentage of emails composed with its AI tools has increased fivefold in the past year.

Is this really a good metric to aim for? Don't we want productivity tooling to result in less email not more?

I find this particularly fascinating, given the post email world I live in now. I haven't had an email from a contact in over two years. It sounds like a sales tool, in a world where the goal is to distance from that availability.

Was this “post email world” a choice or something that happened? How do you communicate rather than by email now?

In my country, e-mail is now used purely as a sales tool. Notifications from government and schools are also delivered via SMS, and either WhatsApp or Telegram. So.. yeah, block everyone.

Turns out the second you do this you eliminate 100% of the spam in your life. Honestly, if I ever lived in North America again, I think I'd also just stop reading e-mail.

You're talking about personal email, they're talking about professional email.

Please do not assume that professional email is used differently. Plenty of companies here simply black hole inbound email unless it's from the same domain or known applications. Frankly, it's refreshing.

> percentage of emails composed with its AI tools has increased fivefold in the past year.

read: spam has increased 5 fold!

“Increased volume of email” sounds like something people would pay to avoid.

That's how companies like Slack get billion-dollar valuations. The promise of "less".

There are many people whose job revolves around churning through emails (sales leads, recruiters, etc). This is a huge win for them.

A better metric would be: How frequently does the recipients of those emails need to reach out for clarification. The goal of any writing should be to increase clarity and ensure that your message is clearly received.

Why do their customers even need to send 72% more email?

Why? The amount of work won't shrink just because you can execute each task faster.

I think Superhuman's CEO in an interview said their product is specifically catered to people who are seeking inbox zero.

For those people this would be a great outcome. The question is should this be the goal of most people? Probably not. But most people are not their ideal customer. They explained their ideal customer in depth in an episode of the Acquired Podcast.

Anecdotally: The Superhuman users I've worked with start skimming e-mails and sending super-short replies. Sending a "Good job team" or questioning something in sentence 2 that would have been answered if they read all the way to sentence 5 of the e-mail is the way to clear their inbox.

The inbox->outbox flow turns into the way to clear the inbox. It's not about better communication, it's about speedrunning their way to inbox zero.

The worst case was a person who would respond to everything with a one-sentence question, then respond to the response with another one-sentence question, and repeat all day long. He could turn a brief e-mail into a thread with 15 one-line responses that could have been avoided by spending more than 10 seconds thinking about it.

This old-ish Newport essay comes to mind:

> The knowledge sector’s insistence that productivity is a personal issue seems to have created a so-called “tragedy of the commons” scenario, in which individuals making reasonable decisions for themselves insure a negative group outcome. An office worker’s life is dramatically easier, in the moment, if she can send messages that demand immediate responses from her colleagues, or disseminate requests and tasks to others in an ad-hoc manner. But the cumulative effect of such constant, unstructured communication is cognitively harmful: on the receiving end, the deluge of information and demands makes work unmanageable. There’s little that any one individual can do to fix the problem. A worker might send fewer e-mail requests to others, and become more structured about her work, but she’ll still receive requests from everyone else; meanwhile, if she decides to decrease the amount of time that she spends engaging with this harried digital din, she slows down other people’s work, creating frustration.

https://www.newyorker.com/tech/annals-of-technology/the-rise...

I'm hesitant to call the email-skimming workflow that you mentioned a "reasonable decision," but I think the point still stands about how one person speedrunning their inbox can make everyone else's inboxes that much worse.

> I think the point still stands about how one person speedrunning their inbox can make everyone else's inboxes that much worse.

I think you're looking a little too strictly through that Cal Newport quote.

There's another big problem that isn't external: The people who speedrun their e-mail like this (which isn't every Superhuman user, to be fair) are also harming their own understanding of those e-mails.

From what I've seen in a few people, it turns into a false sense of being productive while they self-sabotage their own communications. Inbox Zero becomes the goal and they think their job is done when those e-mails are all gone.

That's also very true. Most of the people I've interacted with who communicate that way (the "one question per email" thing hits painfully close to home, lol) seem very... maybe "oblivious" is a polite way to say it. I guess I never questioned whether their inbox-speedrunning is why they're so oblivious vs. whether it takes an oblivious kind of person to think that's an effective way to conduct your comms :P

Yeah, if they could increase the quantity without affecting the quality (or improving it), that would be great. But there's a good chance that is not what's happening.

This is a bit surprising. I even didn't expect Grammarly to have the cash, I used to be a paying customer of theirs when I was writing papers, but apparently with AI I don't even need the free Grammarly anymore.

They might have a lot more cash than you'd expect: https://www.grammarly.com/blog/company/grammarly-announces-g...

and superhuman might be worth less than they raised at - https://x.com/pitdesi/status/1940079704423506401 (or not, we simply do not know)

The undisclosed acquisition price/terms implies a down-round exit at best, and a fire sale at worst.

Yeah localllms replace that job very nicely.

They produce convincing text but the grammar is not actually that good. A lot of missing commas, for example.

I would generally expect llms to mirror how the average human writes vs the most technically correct firm, so that would track.

Which ones? Latest MistralSmall , Gemma or even Qwen3 does really good job with grammar.

That was my thought as well, but one of my college age kids still likes Grammarly over just using ChatGPT for grammar checking and rewriting, says it does a better job.

Excited to see what they are doing now after the "acquisition" of Coda (seems like a bit of a reverse acquisition or acquihire since they buy Coda and have Coda leadership take over Gramarly.

I would actually agree with that, too. Grammarly certainly isn't perfect (it still occasionally struggles with the nuance of some idioms or proper nouns), but it does better than LLMs (I use MacWhisper with its "CleanUp" AI prompt for dictation). But Grammarly's inline use is actually pretty handy (even in this text box, I pause typing for a moment, and there's a Tab prompt that will auto-edit my text live).

Maybe someone else can figure this one out better than me - it is interesting Grammarly raised $1bn of debt, basically. And now they're using that cash to buy revenue-generating companies, I assume at a knock-down price as others have speculated. I assume this is one big strategic play on both sides where each need the other for various reasons.

> Superhuman valued at $825 million in 2021, $35 million annual revenue

This is nuts! I used Superhuman for about a year. And honestly, I might still be using it if the pricing weren't so off. It had a couple of nice features, and the keyboard-driven approach was a welcome change for mail clients.

But ultimately, Superhuman had nothing that couldn't be replicated in a relatively short amount of time (maybe even with plugins?).

$825 million? Maybe I should start a mail client company...

Zoom was worth around $125 billion at the start of July 2021. It’s now worth about $23 billion.

So by that logic, Superhuman may be worth around $165 million.

More interestingly though, let’s assume they spent the $110 million they raised. That means that each of the ~85k customers they would appear to have based on the estimated revenue cost them about $1300. Though probably more as a proportion of ongoing revenue will obviously be driving sales and retention.

I did see something somewhere saying that they have very high customer retention. That matches my anecdotal experience - I’ve been using it for several years as have several people I referred.

But yeah… an $800m+ valuation? That feels like Covid-era hype.

bang on. It is quite straightforward to replicate the Superhuman experience with a plugin - www.simplehuman.email is one

simpl.fyi is the other

Grammarly is one of the tools I pay for, and I am worried about the security risks of using it. Really wish there was an alternative that: 1) Does local processing (local LLM?) instead of sending all my data to their server. 2) Had a lightweight Chrome extension that didn't inject many MBs of scripts on each page.

Feels like that'd be trivial to build, biggest issue is having to ship large files (LLM weights), but maybe CNNs would be enough, I'm guessing Grammarly started with CNNs or similar?

What are you using Grammarly for, is it just spell/grammar checking or something more? Is the UX particularly good? Personally I tried it some years ago but didn't understand/see what is/was special about it.

I personally used to have a subscription for grammar checking, especially for longer papers. Now, I just use a LLM. I personally don’t see the strategic value of them pivoting to using genAI; there is no way I would pay $30 a month for something that will take at most 100k tokens using other LLMs. They seemed to have heavily downplayed their unique aspect which is the deterministic ruleset.

Harper checks a lot of your boxes and is getting better all the time: https://writewithharper.com/

How does Harper compare with LanguageTool. We use a privately hosted version. It's better than nothing, but in practice it's more like a super-charged spell checker.

On the homepage it has a comparison of figures, presumably indicating response time, though it doesn't speak to its performance in terms of grammatical errors caught:

Harper - 10 ms LanguageTool - 650 ms Grammarly - 4000 ms

For a more-classic, more-human experience (i.e., computer flags potential issues, you decide and correct if necessary) there are proselint and vale.sh.

https://github.com/amperser/proselint

https://vale.sh/

FWIW: latest Chrome ships built-in AI APIs

https://developer.chrome.com/docs/ai/prompt-api

so it should be a matter of time to have a replacement extension using this local API. However the built-in model is Gemini Nano.

yeah i highly doubt people will use Nano for more than simple retitling because we're so used to higher intelligence for ~free elsewhere

Superhuman, the most super-email client experience, that only people in the bay area (and some folks in NYC) actually use.

How much did they pay for this? I hope not much.

Those are some of the best customers to have in case your business is under threat and you might need a bailout (M&A) in the future.

You would need close to 90k customers to get $35M revenue as they claim.

90 thousand customers sounds like a whole lot of users to me.

I use it myself and it is by far the best email experience ever created. Is it worth the money? That depends on your needs and work, I guess. CEOs laugh at the cost. Developers might think the price is nuts.

This feels like a pattern of grammarly becoming a holdco / following the Salesforce playbook. Find companies with - solid but not breakout growth (and probably slowing) - a loyal cult following - raised at too high a valuation in the peak era - talented teams - still founder led by strong product thinkers

Salesforce did this with Quip, Slack, etc.

It's not a bad play, like PE rollups but arguably with more potential for success from a "founder/startup friendly" perspective. PE, historically, is very good at stripping/optimizing/etc (think Bending Spoons), but I think there is magic in startups PE simply does not have that allows for a different, yet more successful, outcome (growing vs cutting your way to financial success).

We should see more of this as large, profitable startups have grown into long term private companies with no need to go public.

> large, profitable startups have grown into long term private companies with no need to go public

which i think is a real problem - it prevents "mom and dad" investors from partaking in the wealth creation process, as they are not sophisticated investors and thus barred from being able to invest in these PE investments.

Public listing has become a cashing out operation, rather than a fund-raising operation, if this continues to happen more and more. And the public becomes the bag holders.

I, for one, think it’s hilarious that a company that put so much energy into being ultra-exclusive ended up getting acquired by a company with such a mainstream reputation. Grammarly’s target audience is the people who couldn’t join Superhuman in a timely fashion.

Superhuman made one of my accounts wait for four years for an invite.

Never used of Grammarly and only heard of it in some random feed once a year in the past few years. Is money slushing like water in earth or what?

Are you a native English speaker?

As English as a non primary language speakers globally use it heavily.

English is the default business language, so if you are from another country and are not confident in your English, your gonna install such a tool.

Having applied to YC this round with an email product, this is very interesting.

Almost all email clients, apps and web, have keyboard shortcuts built-in. That being said, I never understood the appeal of Superhuman, except for being… expensive.

I really don't see the fit

https://www.youtube.com/@superhumman

Hopefully this brings some of the Superhuman product magic to Grammarly. Although both products could improve AI functionality significantly IMO.

Oh no. I love superhuman :( Please don’t change it

Looks like Superhuman investors wanted their ROI and Grammarly cant grow on its own that it needs to buy customers.

The founder of Superhuman has really got this playbook down. I remember his previous startup (Rapportive?) was another email-improving thing that got acquired by LinkedIn fairly quickly and got shut down. Not a bad gig.

This was not very quick.

I first heard about it last week, so it felt quick to me.

My history with email clients being acquired is not encouraging. The history is they are effectively abandoned and/or shut down on the order of weeks and months not years. See Dropbox/Mailbox.

I miss waiting in that large invite queue until you were finally let in, what felt like the first inbox zero proponents and possibly(?) introducing the swiping rows with different actions depending how much you swiped.

And, you could reorder the emails inside your inbox!

> My history with email clients being acquired is not encouraging.

Ignoring the constraint of "email client", has there been any acquisitions where the acquired product got better post-acquisition?

I can think of countless examples where it got worse, but from the top of my head, I can only think of maybe YouTube, but then only in the initial post-purchase period, and same goes for a bunch of other examples. They seem to eventually always turn sour.

Maybe GitHub? But it traded "no new features - no downtime" for "some new features - a lot of downtime" after the Microsoft purchase, so I guess it's very subjective, probably at least some people like that tradeoff.

Honestly...true. Even Rapportive (same founder) effectively died post acquisition by linkedin, no?

YouTube. Android. Google maps.

Arm

Diamond Aircraft

Volvo

Cirrus

All retained their culture and brand and the products keep improving incrementally. Parent companies keep a low profile wrt product.

Volvo owners I know would disagree vehemently

barely off topic, but your comment reminds me of SAAB (which i had and loved) being acquired by GM (which phased it out, after making it bland).

> See Dropbox/Mailbox.

And Google/Sparrow.

Really miss Sparrow! To me it was the perfect email client.

Sparrow

> Grammarly's acquisition of Superhuman follows its recent $1 billion funding from General Catalyst, which gives it dry powder to create a collection of AI-powered workplace tools.

Dry powder to do what?! Is this americanism? I've been here for over 8 years and every month I find some wording that's just bizarre, like as if there was a competition for ways in how to confuse someone.

“Dry powder” is cash on hand to fund acquisitions. Yes, it is one of the worst businessisms.

Dry powder as in gunpowder.

Perhaps you are familiar with similar phrases “ammunition” and “war chest”?

You mean actual objects/entities or other financial terms?

As figures of speech.

> Dry powder to do what?! Is this americanism?

No, is financialism, which is basically the same thing but I digress.

> Dry powder is a slang term referring to marketable securities that are highly liquid and considered cash-like - https://www.investopedia.com/terms/d/drypowder.asp

Finance seems particularly filled with terms with no real connection attached to the concepts at hand, not sure why. Sounds cool I suppose?

I hate meaningless jargon as much as the next person, but I think the analogy here is pretty clear and useful.

Dry gun powder is ready to use, allowing you to fire whenever. Much like liquid assets, ready whenever.

If your powder gets wet, it'll take you a good amount of drying before it's of any use. Much like illiquid assets, would take some time to be useful, but still useful nonetheless.

I first heard it a few months ago when Chrystia Freeland, then Canada's Finance Minister, resigned and used it in her resignation letter. The meaning was immediately clear to me.

> "That means keeping our fiscal powder dry today, so we have the reserves we may need for a coming tariff war," Freeland wrote.

https://ici.radio-canada.ca/rci/en/news/2127174/finance-mini...

I don't know if it's uniquely American but I agree it is a very annoying term. It just means cash. Finance bros like to pretend they are "going into battle" or something when they go negotiate an acquisition, so the analogy is that you have a large store of gunpowder ready to deploy at a moment's notice when you want to go to war (go acquire a company).

On the plus side, when you're a seller, you want buyers who have this attitude. That they are going to war and the only way they can emerge victorious is by deploying an enormous amount in your direction. You want buyers who will brag to their friends about how much they spent.

I interviewed at Superhuman almost a decade ago, and the founder did indeed brag [repeatedly] to me about how they bought the superhuman.com domain for $300k.

From the context I figured this was more a "pancake mix" as they have the powder and now they just need to add water. You know, boyz be cookin'. Wouldn't figure out that this is some "war" analogy.

it comes from "keep your powder dry" (ready) in the military sense, I think.

Financialisms like dry powder are fairly mainstream and I'm just glad that they didn't use consultantspeak.

> Dry powder to do what?

This is a very common term for business people and especially investors and startups. It's a short phrase that carries a lot of meaning and packs a lot of punch.

Google Gemini:

> In finance, dry powder refers to readily available cash or liquid assets that a company, investor, or fund manager holds in reserve for future investments or to meet obligations during economic downturns. It's a metaphor, originating from the need to keep gunpowder dry for use in battle, symbolizing preparedness and flexibility in financial contexts.

> In essence, dry powder is unspent capital waiting to be deployed

This one term packs in all of that meaning into two words, so it is quite a useful tool. "Cash" alone doesn't have an implied context, whereas "dry powder" is immediately understood for strategy and positioning.

It's also nice to have analogies that are striking and evocative. It makes language fun and flowery instead of dusty and spartan. Business people have to business all day, and this injects a little flavor and excitement with wordplay. Drawing up images of 17th century battles is nice when the reality is emailing back and forth.

There lots of other phrases like this that you'll stumble upon. Someone should make a dictionary of these at some point.

I guess that explains the Coda acquisition.

Honestly, I wonder what makes anyone think that Grammarly is the right centerpiece for an AI rollup?

So many unexpected moves lately. AI is really changing how companies assess their trajectory and business plans.

They raised $1 billion and bought a small company.

If Ai was so good, they wouldn't have needed to buy a company. They would have vibe coded it in 10 minutes.

They bought the high-value users (CEOs, VCs, etc.). They can then sell to those folks, and perhaps train their tools based off of the way that those folks write/behave (if their privacy policy allows it).

Reply on news.ycombinator.com