still working on https://hcker.news, which has an absurd number of features that improve your QoL when reading hn.
i've massively improved a bunch of things like the AI filter, which now gives you the option of filtering out github repos with AI authorship.
Also improved comments, which I'm serving through my own backend which has made loading of comments super fast, and it's going to be the foundation for some really great other features coming soon.
Soon: HN feature parity via browser extension and sync'd accounts.
Gonna give this the biggest compliment and highest honor I give alternative hacker news interfaces: I don't hate it!
I really like the simplicity of HN and this kinda keeps it (somewhat) in a way that I like.
When the comment overlay is open however, I was expecting tapping outside of it to close the overlay btw, not to let me tap the underlying stories and get redirected.
You're mixing up the numbers. Their annual run rate is $2.4 billion. Revenue grew 140% YoY. That's an 8x sales multiple on good growth. The valuation is not egregious.
Sorry I meant profit. On a 5% interest, you get 1bn (pure profit with no risks) per year for a 20bn of capital. Their revenue grew 140% YoY but does that account for new acquisitions? Also, their profit needs to grow x10 in order to match bonds. It may have made sense in a 0% interest rate world but not at 5.
It's a business model that's like a shark: perpetually swimming and eating or it's dying. That's how they can show big increases in revenue, but the profits are always decaying along with the products.
You missed filmic. Wow. So these people are the reason why Filmic went overnight from one of my favorite iOS apps to something for the trash heap.
my knee jerk reaction is to throw shade at the ppl operating the company but, upon second thought, there's an obvious pattern of them relieving the company from people who knew less how to run (and sustain) it. I haven't used evernote in almost a decade but it actually seems.. fine? I stopped using it when the company started selling merch as a latch ditch effort to make money.
They're basically the retirement home for once-good apps and services who still serve a dwindling core audience but are not longer growing or even a real contender in their field.
At least Evernote was saved by Bending Spoons. At one point, even Evernote was getting roughly a third of its monthly revenue from merchandise, which is pretty wild for a paperless note-taking app and a decent sign that the core business was already in bad shape. For the rest, though, they seem very good at squeezing hard whatever is left.
Yep, but now they’re jacking the pricing to the moon and everyone is starting to leave for other apps. I almost did it this year and probably will do it next year.
I left last month. $250 a year or something crazy like that. Obsidian has a free web clipper, I'm planning on using that since I was just bookmarking stuff.
Exactly. I’m trying out Obsidian this year. In Bending Spoon’s defense, they have been improving the app, particularly the sync, but I don’t care about many of the new features (e.g., AI stuff).
I built a filter for AI (inclusive and exclusive) for hcker.news a while back and continue to maintain it. It’s quite robust, and does more than just keyword filtering.
Never thought to check codebases for AI commits. That's very interesting. Since your comment I have added an option to filter out github submissions that contain commits from an AI assistant. I've implemented it by pulling the last 50 commits from the repo and checking for signatures of the commonly known coding assistants.
Regarding your second one, I have a pretty extensive static list that missed out on "GenAI" as a keyword but I've now created an automation to check search for exceptions daily.
Using dependency cooldowns is not a free-rider problem. There's a real tradeoff here – ppl are trading their time preference for security.
Just as users are incentivized to avoid malware, researchers and attackers are equally motivated to be the first to discover it.
The concern trolling around widespread dependency cooldowns doesn't make sense. Most people shouldn't be eager to download a release that hasn't made its way through at least some scans.
Had no idea Vite and OXC were made by the same company. Makes so much sense.
I don’t get the complaining about OS developers behind these incredible pieces of software like uv, bun, etc is a bad thing. If anything, it’ll continue to incentivize great developers to fill in the blanks and continue to push things forward. It’s a win for everyone.
OXC predates VoidZero and is made by Boshen. Evan had to try for a while until he was able to convince Boshen to join them. OXC is the best of the JS toolchains implemented in Rust, so it was definitely a scoop.
I've seen other parents create AI videos of their toddlers being visited at night by Santa. I've seen parents happily throw their children into AI video generators to entertain them.
People are using AI recklessly. I can't imagine stealing the gift of a child's imagination away from them and instead, replacing it with these hollow representations of reality. It disgusts me.
I use AI all the time for coding, but I've drawn a hard line at the point of intermediation with others.
I think I've spent more on dropbox, lifetime, than most other subscriptions (it's also the first service i thought was worth paying a subscription for). I still pay for it. Drew built a great service.
On the other hand, I can't think of a single new feature they've introduced since 2011 that matters. All I care about is packrat and good syncing. Is there anybody that loves anything they've built in the last fifteen years? I feel like the company could have had a skeleton crew keeping the lights on and I wouldn't have noticed a thing.
Now, in 2026, all I want is for my coding agent to be able to grep the files in dropbox. Feel like dropbox will sooner rely on selling merch than offer something useful like that, though.
>On the other hand, I can't think of a single new feature they've introduced since 2011 that matters.
Honestly that's what I love about it. I work on something on my desktop. Then when I go to my laptop, everything is there too. It's great. When I get another computer I can just enable Dropbox, walk away, and all my projects, notes, pictures, etc. will be there. I pay them some amount of money per month and it just works and I very rarely need to visit the website or even click on the icon in my toolbar.
Sometimes I read notes on my phone and it's kind of annoying that I can't search through text using their app, but I generally consider that to be a problem with Android rather than Dropbox.
I generally have not thought about how Dropbox spent its money until I visited the web interface, which has been redesigned for the tenth time over, and remembered that there’s still no way to see how much space your folders are taking up.
I guess I already know roughly how much space they're taking up since I just check how much space I'm using in my dropbox directory on one of my computers. From my perspective, Dropbox basically has no User Interface, but a fantastic User Experience.
For me that and end-to-end encryption (I know it's supported for teams now).
Instead they just added more annoyances over time. Every time I logged into the web interface, I would get stupid upselling advertisements (maybe don't badger your paying users with that nonsense)? I replaced the official client by Maestral years ago, because they switched to embedding a web browser, and the AFAIR the client was also trying to do upsells.
My wife were and I were customers for years. But we finally decided to terminate our subscription last year. Mostly because of the constant upgrade nagging + the orange guy taking office and Dropbox not providing E2E encryption on family accounts. So we switched to Proton Drive. It's worse in many ways, but at least it's E2E encrypted and doesn't shove upgrading ads in our faces all the time.
E2E is supported for specific types of folders available only to teams but the admin has to enable it and that folder has to be used. You can't apply it team wide to all users. It's a very poor implementation.
Using `age` or `rage` is deadly easy. Also, e2e is most effective when it happens out of band - if the idea is "I don't trust Dropbox, so I want client side encryption" then you shouldn't trust them to do the e2e anyways. I realize it can be more complex, but managing it yourself gives you the maximal benefits of e2e.
It's not just "better", it's the only approach that addresses the real threat model. So you can go with something simpler like checking a box but then you aren't getting the same feature.
FWIW this is the reason, to my understanding, that Dropbox has always been reluctant to support this feature - because if you actually want E2E you probably want it out of band.
Why not just use VeraCrypt containers with DropBox until this man of colour that you are scared of leaves their elected office? That way, your files are E2E encrypted via FOSS tooling.
There is a lot to be said for staying small and doing one thing really, really well. Any time a service I like tries to expand their business, usually to appease investors, that's when things start to go downhill and I start looking for alternatives.
> There is a lot to be said for staying small and doing one thing really, really well.
Man. 1Password is another example of this. They've chased growth and no longer seem to be able to build a browser extension that actually works. I've been seriously considering dropping 1PW because of it.
Funny you bring it up. I was a 1Password user for 18 years and just migrated off it last month when they announced the price hike. It has only been working 50% of the time, and the lack of a reliable auto-fill is a security risk… yet they are raising the price so all their Avenger backers can get a return on their investment. A bunch of movie stars investing in a password manager feels like a huge red flag to me.
I ended up moving to Apple passwords. I really wanted to keep my password manager platform agnostic, but if I’m being a realist, all my stuff is from Apple, they have a history of long-term support, and so far it’s worked great on all the sites 1Password struggled with. There are features I miss, and I did need to manually validate each imported entry and do some manual cleanup, but I was able to transition relatively quickly with a DB of over 400 accounts.
I didn’t want to go into too much detail if you were thinking a different route, but since you’re thinking of going with Apple, here are some of the rough edges I remember in the transition.
I exported 1Password to a CSV and imported it into Apple Passwords. I had spend some time cleaning up 1PW first, archiving dead accounts, but the export had everything in a flat file, so I had to go through everything in my archive and delete it from Apple PW.
Custom fields and some other special fields aren’t supported in Apple PW, this is where I had to go through one-by-one and make sure I wasn’t missing anything and add those things to the notes area.
Non-password data in 1PW doesn’t fit well into Apple PW, so I used a password protected Apple Notes file for that. Not ideal, but I didn’t use those things too much. For software registrations, I didn’t bother securing those too much and filed them away as txt files, most were really old anyway.
The CSV export made 2FA pretty obvious, so I could set that up with Apple. I want to say it made passkeys obvious too, but I only had 1 and knew what it was.
I cleaned out Apple Passwords before the import, so it only had app passwords for login-with-Apple. The way they do this leave much to be desired, to the point that I don’t want to use that feature anymore.
After the import I also found that it tried to be smart about handling multiple entries for the same site, or tried to intelligently name stuff based on the website. It got most of this right, but I still needed to clean some things up and do a little investigation to figure out what a few things were and fix them. Some of this was due to importing archived stuff. I did have to go back to 1Password to make sense of some of it.
All-in-all, I did it all over a long weekend. I kept 1PW around for a month to see if I’d need it and I never did.
In 2016 they introduced document scanning, and I've used that a lot to digitize old papers, lecture notes, all kinds of things. It works well and has a good UI for tweaking corner locations and other things.
That’s also what I care about, but the atrocious client (and the m1 thing) and the constant nagging in the web interface was too much. I cancelled and now use a mixture of icloud, airdrop, and rsync/sftp with remote servers.
Have you tried syncthing? A few years ago I replaced Dropbox with that on all my computers and it's been great. It works a lot like old school Dropbox, except without the cloud-hosted storage parts. It does p2p sync of directories between computers.
I have, but for my workflow, I found that I don’t need realtime sync that much. I either work locally or the source of truth is online. What I really needed was backup (restic and various online storage) and file transfer (rsync and unison works great).
Seems like tacit acknowledgment that IBM mothership is not the right place for a speculative growth play from both a management and capital perspective.
I’m not IBMologist but I do remember how IBM pushed Watson when it was clear that upper management had no idea what Watson actually was. Regardless of the viability of the underlying technology, it’s best to keep such things away from the consultants.
Also, article is very difficult to read. Bad typeface, spacing, coherence and prose. I found the press release less strained.
It's not that the cheating stayed with IBM: Ray Dalio hired David Ferrucci out of Watson to try to make an AI for Bridgewater. The pitch was to make it a very accurate people assesor, but in practice the goal was to tell you who agreed the most with Ray Dalio. The team spent years of their lives taking Bridgewater's money, building basically linear regression on questionnaires, and calling it advanced AI on interviews. It's all documented in The Fund.
Eh, Watson was a classic open domain QA system originally, no deep learning or much of what we think of in an "AI platform" today. It was one of a bunch of such systems that were built in that early 2000s period. They all failed because the approach fundamentally didn't work very well.
> Seems like tacit acknowledgment that IBM mothership is not the right place for a speculative growth play from both a management and capital perspective.
I'm not understanding your logic, can you explain?
What I see with the program and amounts companies were awarded is some level of acknowledgment of the current state of quantum research (i.e. IBM is generally considered the leader) and their pragmatic approach that piggy-backs on current technologies (for obvious speed+cost benefits).
You must not talk to competent people. IBM is very experienced at this grift. I remember when I used to go to conferences in a different field and IBM would announce "state of the art" results that were very obviously done by cheating (making an ensemble model and tuning the weights on the test set). Everyone doing real work would ignore them, and then they'd go sell to clueless midcap companies on the basis of that announcement.
> I’m not IBMologist but I do remember how IBM pushed Watson when it was clear that upper management had no idea what Watson actually was. Regardless of the viability of the underlying technology.
So pretty much like any other AI company in 2026 hunting for VC money?
i've massively improved a bunch of things like the AI filter, which now gives you the option of filtering out github repos with AI authorship.
Also improved comments, which I'm serving through my own backend which has made loading of comments super fast, and it's going to be the foundation for some really great other features coming soon.
Soon: HN feature parity via browser extension and sync'd accounts.
reply