They make their money from enterprise software, for which privacy and access controls are very important. The other tech leaders on the board are more consumer-focused businesses, and consumer-focused businesses generally don't care as much about privacy.
You're 100% correct and the replies to your comment are missing the point. People can have whatever complaints they want about the equality of encryption etc (and maybe those are valid, I don't know), but your point still stands. People here seem to be missing that when you're dealing with enterprise customers, their primary concerns are security, compliance, etc. So yes, running a company that sells to enterprise gives you a very different perspective from what you'd have running a consumer media company. Whether that helps FB I have no idea, but that's a separate debate.
How it helps FB is that it is trying to expand its market into business software - aka using Facebook Groups and FB Messenger, on your own shard. It competes with Slack and other similar worksplace IM systems. Drew Houston's experience at Dropbox is greatly helpful as Facebook tries pushing harder into that market.
AFAIK they don't, though. Dropbox is still primarily getting its revenue from consumers, not b2b. They have been attempting to pivot to serving more enterprise customers, but are struggling to get traction there.
Given that, combined with the second part of your comment, kinda makes the parent comment's point.
1) Files are not encrypted on Dropbox servers. 2) The Dropbox client scans files on your machine that are outside the Dropbox folder. 3) collaboration with others via 3rd party app requires you to grant the app access to your entire Dropbox directory.
Not only is privacy not a focus at Dropbox. It isn't even a consideration.
Claim 1 is 100% false also (ex Dropboxer here), they encrypt files at rest in 4MB chunks for all versions of the product (from free, pro, plus, business and enterprise). It’s in their security white paper too: https://www.dropbox.com/static/business/resources/Security_W...
Security and privacy are not the same thing - and how the files are stored on disk is irrelevant. Obviously the files are also encrypted in transit as well. The bottom line is Dropbox holds the keys on their end and has unencrypted access to the files on while on their servers. They can read the files at will. That is not an architecture that is focused on privacy - which is fine, but it shouldn't claim to be otherwise.
Most people in that thread are just complaining of high disk usage without specifying where access occurs.
Then page 3 has the following user report:
> I'm running Windows 10 and my Dropbox is located on my E drive (E:\Dropbox). I also notice high disk activity. However when I look in Resource Monitor I noticed that the high activity generated by Dropbox.exe is on the C drive (not E).
To which an employee says the reason is that dropbox is installed on C:.
> Although you have your Dropbox folder on your E: drive, the application is still installed on your C: drive, so that’s why you’re seeing the activity there.
Sorry, I did a poor job explaining. The smoking gun is CPU usage not file system activity. Perhaps you know more about this and can explain it to me. If I copy files from one place on my system that is outside of my Dropbox folder to any other location that is also outside of my Dropbox folder, how is it even possible for the Dropbox client to consume CPU resources if it isn't monitoring files outside of the Dropbox folder?
Here is link showing more people reporting the same problem. Files being copied from one part of the file system to another and the Dropbox client consumes 60%-100% CPU. I cant even comprehend how that is possible.
> how is it even possible for the Dropbox client to consume CPU resources if it isn't monitoring files outside of the Dropbox folder?
I did not debug this or look at anything, so I can't tell you.
One plausible explanation is they are effectively polling Dropbox folder instead of blocking until actual changes occurr. So maybe that hogs CPU scanning the Dropbox folder at inappropriate times, meaning that other activity on the machine has to compete for CPU time and lock acquisitions with unrelated but very noisy requests from the Dropbox client.
The fact that some users say that adjusting ACLs might get their client "unstuck" would fit this theory. Maybe they had a bug that generates noisy scans on the disk when they get "access denied" in an unexpected place.
Dropbox supports two types of 3rd party apps each with different file permissions. 1) Sandbox, and 2) Full Access. If you jail by registering a "Sandbox App", you give up the ability to share/collaborate in a directory with other Dropbox users. To have a third party project that provides sharing you must abandon the "Sandbox" permissions and instead request "Full Access" - which means you get a token that give you access to the users entire dropbox directory. I have several email threads with the Dropbox team discussion this issue at length. Perhaps this has changed since I was neck deep in this issue but I don't see anything in their changelog to suggest that.
You seem to be living in some imaginary world in which Dropbox is able to sell to enterprise companies of all sizes and yet privacy is not even a consideration. A little weird how most companies are extremely hesitant due to concerns about compliance, regulatory and privacy issues and yet unbeknownst to them (but known to you apparently), privacy isn't even a consideration at DB! You might have legitimate complaints about the level of security DB offers, but to claim privacy isn't even a consideration is just stretching it to absurdity because DB would have ZERO chance or selling to a large portion of its highest paying customers if privacy and security weren't considerations at all.
I haven't made any complaints about Dropbox's security. I made complaints about their privacy - which you are incredibly naive about.
Have you ever built a business on the dropbox API? Have you ever had to store and manage thousands of "full access" tokens of other users? Are you aware that every 3rd party that has a "full access" token has unbound access to read any of the files since the files are unencrypted?
The following url tells you how many 3rd parties have "full access" to your dropbox files. Take a moment to appreciate your privacy is dependant on the security and discretion of each and every one of those 3rd parties.
Excuse me for saying I think there is a lot of unnecessary rage-caps and hostile sarcasm in this reply. Gp is not personally attacking you, but it reads like you take threats to Dropbox's honor pretty personally.
I do not think gp is living in imaginary worlds for example.
Dropbox might be secure but I wouldn’t say they are privacy focused. There are other offerings like Tresorit or Spideroak that offer end to end encryption of files, that’s more something that I’d call privacy focused.
grellas' comment has 270+ replies, and HN's software currently paginates threads at 250 comments. That is purely for performance reasons, and I can't wait to be able to turn it off. If you go to the subsequent pages, there are many other top-level comments:
grellas' comment is at the top because the ranking algorithm put it there; many of his comments have gotten upvoted heavily over the years. Moderators didn't do any of that.
People here don't like to hear things that smell like politics, but the Condi Rice thing is a big deal. Excuse my bluntness, but she was one of the public faces of thousands of preventable deaths in Iraq. To have no problem with this I think has so something to do with what people mean when they say many in tech have basic ethics problems.
For example, in ETL pipelines, I would greatly prefer to have an entire DAG go down quickly and noisily than to risk having it generate incorrect data. It's the difference between a crappy morning, and a crappy day or even a crappy week.
Just curious since a lot of my work is on an ETL - what are your favorite DAG libraries/approaches? Some of my ETL workflows can run in parallel with each other because there aren't data dependencies between them, and others should definitely crash loudly
We use a home-grown solution, and try to keep it pretty minimal. For example, we can run all our pipeline steps serially and still get all the work done plenty fast, so, for now, we're keeping that parallelism can of worms firmly shut.
It's also viable because it's the convention. Library calls are likely to return {error, {the bad data you sent}} so it's easy to write receive {error, _) -> {missiles, {launch false}} end.
And, the fact that an Erlang server is explicitly intended for handling zillions of tiny, independent requests. A request fails? Meh. They'll call again if it's important. What matters now is the next tiny, independent request.
I imagine so. I think my question (maybe I missed the mark) was is that does as much as that leak into HN (where I consider the level of discourse far higher than the Internet as a whole) as much for other topics?
afaik some explain this with resentment for not having seen the signs of a good investment before time, considering all the wits that should float around here.
Tangentially related, but I have stopped taking pictures on hikes nowadays. I just listen to an instrumental piece of music a couple times and when I play it back a couple weeks or months later, I am able to vividly recall the experience which I could never do from photos.
I wish we could do the opposite because every time I listen to the cool music I discovered in high school I'm unfortunately teleported back to all the emotional distraught of those years.
I've found that the associations I have with a song do get weaker the more I listen to it in another context, so it may be possible to disassociate those memories by forcing yourself to listen to those songs enough times.
But I've only noticed this with songs that conjure relatively benign associations. Stuff from times and places I really don't want to evoke I also avoid.
> Stuff from times and places I really don't want to evoke I also avoid.
There are songs I can't listen to like before anymore after some conversations (eg lightning crashes by live) or some time (Champagne supernova by oasis).
But likewise, there are places I don't have the energy to hang around anymore.
This sucks. I live with music but when I feel really bad, I no longer listen to music, to avoid tainting it. Now, days when I feel bad are quiet.
Though I mostly enjoy travelling back in time through music, including to high school. Fortunately these years were not traumatising for me. Specific songs are particularly strong for that.
Some songs are strongly associated with specific emotions that were not very enjoyable at the time, but are strangely comforting...ish today. I hope someone gets this and formulates it better / differently. This is probably a kind of nostalgia, but not really in the sense that it was better back in the day.
The first time I ever heard the Numa Numa song was when I was 10 and had the flu. I was puking in the hallway and had intense nausea for hours. My classmates kept playing the song because it was the new viral thing.
It took me over 10 years before I could listen to the song without feeling sick. It's fine now though.
Maybe there are ways to build up tolerance to the music by combining it with activities that do not allow the mood to catch on because they are too incompatible? Like running or working out vigorously while listening to this music? Maybe the link between the music and the memories would wear of with time?
I'm not a therapist but this sounds like something I would try on myself if I were in this situation.
Back in 2005 I was gifted both Guero (by Beck) and Guild Wars on my birthday and ended up listening to the album on repeat while playing the game and I really vivid visual flashbacks to that game's art style whenever I hear that album.
As a child I was gifted an album (cheesy christian rock, but it's what my parents would buy me) and a full set of the Chronicles of Narnia. Album went on repeat while I read the first few books and now they are forever, deeply linked in my memory.
DOOM had Smashing Pumpkins in it first: the cheat IDSPISPOPD ("smashing pumpkins into small piles of putrid debris") gives you the chainsaw due to some inside joke. The Smashing Pumpkins sample is a nod to this.
It's fun deciding to imprint yourself in that way, it feels like a conscious decision to re-wire. That said, the most deeply-seated ones are the ones I didn't really decide on.
along the same lines, red hot chili peppers and halo 2 on Xbox live for me!
also morrowind since somehow I managed to accrue 250+ hours of play time on that despite spending about 16 hours out of every day that I wasn't at school (those days were more like 8-10 hours ಠ_ರೃ) playing halo 2 online.
I've taken to playing certain songs for certain scenarios. My favorite example of such being Blast Off (Human Pyramids) played just as my flight is taking off, that windup taxi and the acceleration to flight.
I've kept that particular song pure specifically because I've captured the essence of traveling/vacation in that 5 minute period.
The problem I have with recalling memories/feelings through msuic, is that the recollection gets weaker every time I listen to it. So there are a number of songs I listened to often to relive a particular moment, but listening to them now only vaguely recalls that moment, and additionally also has the association of reliving the moment later. Reliving the reliving of a moment is less vivid than the initial reliving, unfortunately.
I also have experienced this "musical recall", but I don't think vivid memories completely replace photographs. No matter how well you can recall a memory, there's still the task of communicating that memory to others. Photographs and videos greatly help with this.
No, but photographs can destroy vivid memories, or at least commingle and taint in ways that make them impossible to distinguish from the original memory. I avoid looking at photographs of events I've personally experienced. Taking photographs is disruptive in a different way, but it's difficult to take photographs and not look at them later.
If anything, our brains taint and alter memories over time, which causes them to be misaligned with photographs. The photograph doesn't lie. The brain lies to us and often.
Sure, memories fade and become distorted. But that's precisely the problem with photographs: they're intensely vivid. So every time you look at a photograph you're creating, transforming and reinforcing an experience you never actually had, accelerating the loss of aspects of your personal experience.[1] Photographs aren't the only way to do this: simply hearing people recounting stories can do this, but photographs are just an immensely powerful and almost instantaneous way to have such vicarious experiences.
I'd rather have my own memories, however faded and worn, though there are strategies (like music association) that can help minimize that. Nonetheless, however distorted, they're principally distorted by my own emotions and desires; they're faded by my own lack of interest; they're mine, which I'd argue is the very definition of a personal memory.
Many people couldn't care less. And that's fine. And if you don't know (or believe) this happens to you, then I suppose it doesn't matter, either.
[1] I want to say that losing the personal visual memory might be related to losing the memory of other aspects, like the emotional experience, but that's more of a fear than anything I'm confident I've experienced personally or have read in the literature.
You're saying looking at a photograph is like overfitting your brain to a specific aspect of a memory. I don't see how music association is any different in this regard. Looking at a photograph of a memory represents the memory in a certain basis. Listening to a song does the same thing, just on a different basis. So this idea of overfitting or potentially losing part of the original memory is not unique to photographs.
I would guess that having as many associations as possible is the healthiest for any memory. This is also the principle behind mind palaces and other memorization techniques.
I have a really strong memory for that kind of stuff yet minor details like what time my flight is I have to re-check constantly. I can also recall music itself pretty much verbatim regardless of the complexity but not the lyrics. Fingerstyle guitar pieces I heard years ago are still burned into my brain yet I couldn't sing you the same song if it had lyrics. I would just remember the notes.
Entirely baseless speculating here, but I wonder if music as notes, instruments and structure aren't stored in memory more similar to spatial memory which I also find to be rather permanent.
I visited Washington DC on a band trip in high school about 15 years ago. Right before I left, I bought Five For Fighting's CD: The Battle For Everything. I listened to it the entire time I was there (as I only had a CD player). Anytime we were on the bus, I was basically listening to that CD.
Now, If I ever hear any of the tracks from that album, I have a flood of memories that come back, and I think that is the reason why I remember that trip so well.
Works really well with podcasts for me. My memory of daily experiences is generally incredibly poor, when I first experienced nearly photographic recall after re-listening to a podcast I almost cried. It's a cruel joke that this capability is in there but I can't seem to activate it on my own.
I do this with code and architecture, or any complex system really... It works pretty well! It's similar to physical space having a sort of memory... except it's mental space.
I recall reading in 'The Vital Question' (also by Nick Lane) that free radicals (produced when a cell is unable to meet energy production demands) signals the replication of mitochondria. Antioxidants suppress this signal which leads to cell death. Interestingly an increase in free radicals themselves beyond a certain threshold also result in programmed cell death
