Imagine building a company that lets users store files online, seeing chatgpt released, and then laying off 16% of your company to.. use ai to store files better? I really don’t see how new ai capabilities lead to big layoffs and new opportunities for Dropbox.
snark incoming - Imagine thinking you understand the product strategy for a 14 year old, publicly traded company with thousands of employees without working there or having any sort of insider knowledge.
It's also not much of a reach to imagine use cases here. AI trained on a company's corpus of internal documents is an extremely common idea right now. DBX is sitting on a trove of data that is specifically well-suited for building bespoke, company specific AI models.
To add to that, Product leadership at technology companies has evolved to be much less technical, so they aren't going to have any deep insight about AI that, say, a software engineer who has worked in that field might have.
But hey, if in a few years Dropbox launches a popular AI-based feature I'll gladly eat my words because I'll probably be using it. They could benefit from enhancing their search functionality, for example -- but that's just a nice-to-have feature, not something revolutionary.
However, I also totally see the point, as an example: the biggest reason I like Google Photos is the automagic tagging on the pictures, I can just write any random description and it usually finds them across 10+ years of photos I have stored (not perfect, but good enough), I can see AI replicating if not even augmenting such functionalities.
I worked at a similarly sized company to Dropbox, and when I questioned product decisions I would often receive similarly smug put downs. How could you possibly think you know more than the CEO?
And then I watched over the next few years as the product strategy floundered, reorging many times without shipping anything concrete.
I suppose what I’m saying is that while your comment shuts down conversation quite effectively, it can still be wrong. Sometimes - quite frequently, in fact! - there is no man behind the curtain. Sometimes, when product decisions seem questionable, they really are questionable.
(Sorry if I have an annoyed tone; I think I'm subconsciously venting against all the times I've heard this from others in the past.)
Snark incoming. Imaging a society so shitty that you have to try to "understand the product strategy for a 14 year old publicly traded company"
That is, one society in which companies dont go about their business, aiming for sustainable growth and a healthy profit, but pivot to whatever unrelated technology (or, more often, bullshit fad) comes next, to prop their stock or to increase their valuation when they're sold.
I've worked at a pretty wide range of tech companies throughout my 15+ year career. When I was young I would ask myself "what is leadership thinking!?", my realization later in my career was simply: "oh, they're not thinking"
14 year old and younger tech companies have only existed during a tech boom period, when money flowed easy from both investors and customers. There has been zero market pressure to put thought into building products.
In the case of dropbox in particular, any long term customers (such as myself) can confirm that there has clearly not been a coherent product strategy for at least the last few years.
So we can just point out it's strategies on it's latest 10k filing, where it's investments have been and where it's gets it's money from?
Many of us have worked for long-standing public companies with thousands of employees that have embarked on product strategies that were not particularly valid.
Facebook comes to mind, for example, and its Metaverse boondoggle. Or Google and half a dozen things. Evernote, though it's not a public company, also comes to mind...
Did anybody need "insider knowledge" to know Google's Stadia was DOA? For that matter, Dropbox has tried pivoting to mail and apps (Carousel + Mailbox) in the past, and I don't know whether insider knowledge was needed to predict that failure.
I can imagine use cases for AI + Dropbox. Whether they are urgent use cases where customers see a real problem they're willing to throw money at is another question entirely.
> It's also not much of a reach to imagine use cases here. AI trained on a company's corpus of internal documents is an extremely common idea right now. DBX is sitting on a trove of data that is specifically well-suited for building bespoke, company specific AI models.
What could go wrong in dropbox invading their customer's privacy to scrap out a few pennies mining their data?
Imagine I'm a shareholder and the product strategy is either bad or being communicated to me unclearly.
> It's also not much of a reach to imagine use cases here.
It's also not much of a reach to imagine the same breathless magical thinking about AI which is happening everywhere taking over the management at Dropbox.
So far I've seen a lot of potential and exactly 0 effective applications of AI at scale. So I know which my money is on.
I'm definitely betting on AI, but the way I'm betting on AI is betting on AI companies building better AIs. I'm not betting on companies whose only involvement in AI is that they suddenly have to have it because it's the latest buzzword.
Snark incoming - if somebody at Dropbox reads this they might have an AI strategy now. I think your idea is amazing. But I think that a lot of publicly traded companies are now "doing something with AI" like a headless chicken without any deeper though on why.
But your idea is good!
They were also non-E2E encrypted for like most or all of those years. It was the main reason I didn't use it for personal backups. I use SpiderOakONE (paid) instead, which is.
So yeah, they have a fuckton of other people's documents. 100% chance. Even I admittedly use one of their free documents as a simple todo list at work. That's probably really nice data for them too.
If they aren't doing some ML based on that data, somebody at the company is not doing their job.
Then you go on saying this is a common strategy and is well suited?
Do you understand the strategy or is your comment as worthless as you suggest in your first paragraph?
Understanding product strategies are part of many users day to day lives. Don't dismiss everyone and put them on your level of understanding.
In other words, encourage this mindset wherever possible, it's going to be what causes the next wave and spike in demand for SWEs.
You’d probably be better off using a different service if it bothers you so much.
ChatGPT can come up with completely made-up answers that just seem right.
Dropbox moved beyond basic file sharing long ago. I don’t personally use all of their productivity tools, but I know a lot of people (generally business/creatives) who do.
If you try to interpret this through the lens of “Dropbox just stores files” then it won’t make sense. However, there are a lot of AI opportunities in file management. Imagine something like an AI-powered search through your documents or an AI photo search. Would be much easier to find that photo of your family standing in front of the Grand Canyon if you could just type that in.
You may not want that product, but it’s unreasonably snarky to pretend that a big company is dumb for pursuing this obvious opportunity.
File syncing and storage is a feature rather than a full product, these days. Microsoft and Google have made sure of that. You need something more to compete. So, it makes sense that Dropbox went in that direction.
But it does smell like most tech bubble in history. Every publicly traded company has to somewhat fit the last au gout du jour tech in their announcement/roadmap/hiring to satisfy shareholder and bump stock prices. Happened with the dot-com bubble, web3, metaverse, cryptocurrencies, "BigData", ...
Although I get that the idea behind is that nobody want to miss "the next big thing", so if you can afford it, might as well invest in it, and if it is a flop, oh well, but if it does end up being the actual next big thing, you might have saved your company.
CIOs love this one weird trick to boost shareholder value!
Can a utility just stay a utility?
Only companies where owner is top boss can, sometimes. The moment boss has to answer to investors, growth is required.
Only small indie utilities.
For most other things, to grow initially to serve the intended functions to as much of the target audience as possible, requires investment. Investors want that growth to continue, even if it makes little or no sense to the original objectives.
Capitalism - regulation => utilities's entire point is to capture a monopoly to then abuse it.
This is an _excellent_ use of AI, in my opinion, because even if the search results are slightly wrong (they mostly aren't), you're OK with it, and you didn't have to do anything yourself except paying for the service.
[0] https://docs.jottacloud.com/en/articles/7262803-ai-powered-p...
I guess they have Paper, but last I used it, it felt like abandonware.
I could see value in using a chat-like interface with dropbox:
> "Give me the budget for the R&D team from 2019"
Google and Microsoft can do much outcompete them with their storage SaaS products.
Also I always hate “to those departing…” as if it was their idea. These emails always promise fully owning the decision but they never deliver.
"Mr. Robot, categorize our customers in a monetizable way by inspecting the data they hold in their accounts."
Maybe they'll come up with something of value to bolt onto what they do today, or maybe they'll add a totally different product to their portfolio, or maybe they'll pivot away from their current business entirely, or maybe they'll just waste a bunch of money chasing the current fad.
I can think of plenty. Text-based image search, "find the latest TPI report last modified by so-and-so", etc.
Which isn't to say that Dropbox won't spend most of its efforts thinking up features that most users plainly don't need or want, and in cranking out buggy versions of these non-features at the expense of their core product. Like all these companies, once they get sufficiently long in the tooth.
Yeah, well good luck finding people with actual AI experience right now because everybody's looking for them. They might have been better off to retrain some of the folks they're going to layoff. Hiring people with these skills can take a while. Maybe they could be bringing some folks up to speed in the meantime?
Now they're creating completely new products to include with the same subscription that have little or nothing to do with the original product.
Document writing and signing, password storage, screen capture, video review/approval process, invoices (beta)...
Imagine selling Rolodex in the post-email era.
How many of the laid off "team" members could have been retrained?
ChatGPT will just generate your documents on the fly from prompts you memorize.
These generated documents are presumably files? And the long list of prompts would ideally be stored somehow, likely in a file.
combined with the biden depression, this makes it a good idea to fire employees and gain this new source of income.
Honestly I think it's actually a pretty fantastic development if that's the direction. I don't use Dropbox much, nor have I used aifiles (due to sending all of your data to OpenAI) - but the idea of being able to not have to manually and tediously look over terabytes of files to completely reorganize all of your files into a better directory hierarchy, and tag each files with meaningful labels, etc sounds phenomenal.
Obviously there are some implementation details for this to be not awful - for example: 1) only local models for local data, 2) making the changes on e.g. ZFS (to allow rollback) or as some type of optional 'overlay' view to switch back and forth from the original to ai-organized, etc. and 3) having thresholds and logic for what may be considered 'duplicates' to be removed, and how to better compress data
As for the de-duplication and processing: this could be very good for dropbox in that, e.g. if a person wants to completely re-encode all of their image files or video files with AV1, the resulting data could be cut in half or more - which saves Dropbox storage space. After which, neural perceptual hashing could be done on all of the files and a threshold of similarity could do de-duplication on a perceptual basis (for example, keep the bigger size file that is 99% similar to a 2x downsized version, and re-encode it). User preference to keep things like tiff files completely intact or any other lossless encoding of their choosing could be good options as well
There's definitely a strange disparity between the computation cost for deploying a decent model to do this compared to the storage cost - but if a (perhaps even non-LLM!) small model is created to be able to plow through data at fast rates could be deployed it may make sense.
Or perhaps the type of semantic compression that LLMs do are of interest for making a new type of lossy compression algorithm of which Dropbox is interested. It is certainly interesting that even the largest LLMs we are aware of contain only a few hundred GBs of weights, but have information on various topics from Pokemon to electric engineering, to cellular biology, to cake decorating - yet one instance of the best compressed pubmed database I can think of (only containing the small fraction of cell bio and med part of the LLM - perhaps 2% of it) is like 60GB with Zstd compressed (yes, yes, I know lossy vs lossless - but still) - which would be probably more like 1/2 or 1/4 of a typical SoTA LLM. So clearly lossy compression via including semantic info is hugely beneficial over typical lossy compression techniques (and perhaps one could argue, what often lossy compression algorithms try to model in the first place, in a very rudimentary way).
I mean I wanted to store some old tax returns, medical bills and crap in DB, AI comes in and decides "bah, this is crap!", delete files silently to make space for more exciting stuff. Everyone is happy now!
Anyway, good luck to the former Dropboxers, I hope they land on their feet.
The idea that some sort of complex algorithm should be used indicates a misunderstanding of what business management actually consists of.
It is much more social science than engineering.
Yes, in this case the corporate goal is to fire 500 people as a sacrifice to the finance gods.
If the goal was 'refocus strategy', you'd probably want to actually assess your business and determine who is no longer necessary, and you would doubtfully end up at a nice round number
> it makes me think that they decided "how many people do we need to sacrifice to appease the finance gods" rather than "how many of these roles are really needed to achieve the goals we set ourselves".
Any big layoff involves, or is possibly the result of, shifting goals. Company goals change, especially as changing economics makes them more or less possible.
Specially important since company was founded and grown to employ thousands of dropboxers without any money from finance gods.
500 is also a businessy-salesy round number, compared to 512 which is what an engineer would have considered a round number. Gives some clues as to where the number came from.
They're just following the trend, or listening to the ridiculous demands of investors following the trend.
> I thought from one of the smaller companies no one would notice.
> Like one of the cab companies.
>> Fire one million.
> But 500,000...
> One million. Fine, sir.
> Sorry to have disturbed you.
On this thread, there’s the following sentiment… “Dropbox has access to your data so they are well positioned to do some AI stuff on your data.” But also you have your data so you could work with really any party you want, and Dropbox has few advantages by being an existing custodian. Custodian status probably is not the differentiating factor in a race to build AI tools.
AI optimism. I just don’t see the argument. I don’t see how it is good for Dropbox, or the enthusiastic pundits, or even me! (I use the tools, too, but my income and quality of life have not skyrocketed.) My suspicion is that things like Dropbox laying off 16% becoming part of the AI optimism narrative – and AI optimism in general – has little to do with the prospects of AI and much to do with the reactive patterns of opinionated storytellers.
Specifically, I think people enthusiastic about AI prospects are enthusiastic because it is the one empowering (encouraging) response they can take… not because it is warranted. That’s not criticism, just an exploration, and I’m well-aware that I could be wrong.
This is a layoff announcement. We shouldn’t expect any AI realism from it.
As a private individual, yes. However if you are a company you might not be legally able to send your data to just any party or it might be a hassle to set up. If you have a contract with Dropbox already…
What has AI got to do with me putting my files in the cloud so I can sync them across my devices?
There is no substitute for being organized.
If the AI can interpret what I mean, find the file, link me to the sources and distill the information and context, it will be major productivity boom.
So while as a personal user you might not see much potential for AI (though I do), I'm sure there's huge opportunities for their business customers to use AI tooling on the many corporate documents and files they already stored on Dropbox. Also if you look at DBX's current offerings you can see they've branched out from just storing files. Signatures (HelloSend), "DocSend", password vaults, etc. Focusing on the potential for AI to solve problems for their business customers makes sense to me.
AI is allowing for massive monitization of what was previously "flat file" style data.
Dropbox, as I mentioned, has always 'not been you friend' - now they are just a more powerful 'not your friend'.
I'm mindful with what I put on cloud storage (mainly photo/video backups with Google Photos). I'm fine bumping up as I use more, but I don't feel like you suddenly deserve to pay for 2TB cause you went 3MB over the 200GB limit. And for how cheap as storage is these days, I think the pricing models should be reconsidered.
I pay for Netflix and I watch less than 2 movies a month. Am I angry because I'm paying for the right to watch multiple movies/shows every day which I will never have time for? No. I get what I want for a price I'm willing to pay.
Every time you buy a piece of software or hardware you get features you don't care about. Do I get angry that these features are included? Of course not. I just don't use them.
But somehow when it comes to quotas people feel like the price should somehow correspond to the cost of delivering the service. But it doesn't. It's just price segmentation, and as long as the pricing tiers are clearly advertised I don't think you have any right to be mad.
They don’t want to sell you a cheap commodity. They want to sell you convenience at higher margin.
I needed about 20 GB of storage. Why the hell should I have to pay for 100 times that?
Since I switched, I've slowly grown to 30 GB, but I still have plenty of room before I need to go to the next tier and pay another $1/month.
Oh well, between Amazon syncing my photos as part of Prime and GDrive I have plenty of cloud storage for my needs.
I think I remember the tier prices being roughly the same since the service came out.
I get that salaries went up which adds to price, but over time it feels like storage price should go towards zero.
The answer is you're not really paying for the storage. You're paying for the servicing. The application with its network connectivity and data transfer, plus the overhead costs of billing and support when necessary.
The proposition may look like "$100 for 10 TB", but that's not a ratio that you can halve or whatever. It's $100 for 10 TB and all the network servicing and support apparatus, which doesn't change if the storage amount does.
It's not that Dropbox couldn't offer a tier for less than 100/year, it's that they don't want to. We around HN know about reducing headaches by avoiding pathologically cheap customers. Dropbox makes more profit overall by getting some of the would-be cheapos to round up to 100/year and letting the rest walk.
When you have less scale, you need to charge a bit more for your initial level.
As someone laid off 2 weeks ago from a company with much less notoriety, I cringe at the idea of having to now compete with 500 of you.
Good luck out there.
As an European, I don't think I will ever get used to the fact that in the US people can be told "go home, you're fired" like that, effective immediately.
It means that I can be more flexible in pursuing opportunities. Businesses also tend to be more risk-seeking because the downside of hiring someone is lower. That benefits me as an employee because it is much easier to get hired.
I’ve been laid off, so I also felt the negative consequences. The key for me was building an emergency fund. Granted, if a person is in an employer's market, the short notice system is terrible.
Replacing the Dropbox client with Maestral https://maestral.app meant a much happier system with a lighter application running. Also no prompts for the SD card to be uploaded.
When Dropbox limited an account to 3 devices, it made sense to use Google Drive more.
Google kept changing the naming and functionality of its client. The final blow, though, is the way that they insist on showing a screen of files instead of the folders that were carefully created.
A lot of people don't use folders very much or at all and that's definitely the trend. But it means that all the work somebody does in organizing their files is thrown out the window.
Google Drive seems like it's more and more forcing people into having a Google Account even if they use a hotmail or yahoo email address.
iCloud with a family subscription means that a family of 5 can have all of their devices and files backed up for $10/month. BackBlaze for laptops and desktops is $7/month these days so $10/month seems like a reasonable deal.
Dropbox decided that the business world was their market and that's reasonable.
For individuals, there still seems to be a market for file serving with no ads, good security and easy access for people regardless of their email domain. Say sharing photos from a family trip or the soccer games your kid played in.
That's a much leaner headcount than what I would expect for a company with as large a market footprint as Dropbox.
Also, what were these 500 people doing anyway? Dropbox has basically been the same product for years.
I'm assuming you're just talking about your personal experience with their consumer product, because this statement is laughably false. Dropbox has made a tons of changes and acquisitions to beef up their enterprise offerings, which is clearly where they've been focused, and where the money is.
>Role Description We are looking for a technical leader to help shape the Machine Learning agenda for Dropbox’s new initiatives. This will involve developing high impact ML solutions that touch hundreds of millions of people and a lot of data. From images to documents in every language, using the full range of Machine Learning techniques from deep learning, interacting with modern LLMs (natural language), computer vision, etc. and executing everywhere from mobile devices to large clusters on our back-end infrastructure.
They pretty much don't know what they don't know and are looking for someone to build it for them.
If I had to hazard a guess at the strategy, they are looking to build an in-house multi-modal model to power enterprise search.
* How many are going to be fired.
* When will employee know they are fired.
* Severance package.
No amount of thought and prayer are going to help, straight to the point.
These transitions are never easy, but I’m determined to ensure that Dropbox is at the forefront of the AI era, just as we were at the forefront of the shift to mobile and the cloud"
Key takeaway, CEO thinks shift to AI is as consequential to his business as the shift from landlines to cell phones Et al.
AI is super cool, but at least for storage of my files in the cloud I prefer something "dumb" to be honest
AI is going to be extremely useful when I can query my files as a knowledge base directly. Like game-changer useful.
The startup I was at recently laid of 15%. I'm pretty sure it wasn't for (public) optics, because the layoff didn't get any press coverage.
And based on the startup's business, I very much doubt the layoff was triggered by ChatGPT's success.
My best guess would be behind-the-scenes VC strategizing, but this is happening even with publicly traded companies like DropBox.
Is there evidence that this is the specific reasoning that companies are using?
Seems like a poor excuse to cut people.
I've brought this up before to their support team. But I guess they aren't interested in small fish accounts and only want 2TB whales signing up.
To be clear: I'm a fan. Just sincerely don't know the benefit of what could be a very targeted and strategic company in having so much staff.
I see this same comment on every single one of these layoff threads and they always seem to be a bit disconnected.
Would anybody else be content with mass layoff messages in JSON format? { severance, healthcare, perks }
I’m okay without the explanations or attempt at empathy thats partially designed for other stakeholders like vendors who want to know the company is healthy
Just take the media L and move on, I bet there are people that would like them for that approach and attract talent. Its pretty insufferable to be around employees that actually like the “family” coddling talk so if they’re not there to begin with then its a better work environment
Feels a bit like having an illustrator create a corporate Memphis graphic of a layoff.
There's also a paper-textured fern in the background.
Hopefully they'll figure this out and rebound from their predicament.
Is every company just planning to feed all of their customer data through GPT / OpenAI, or will they build their own AIs? The first choice seems risky, from a business, legal and privacy perspective. The second one would require a large amount of capital and has many other risks.
I imagine foundation models become a bit like cloud provider paas offerings vs rolling your own data centers.
I feel like HN should just autogenerate these comments now for stories like this (perhaps with a helpful response explaining what the corporation actually does do and giving some examples of corporate jobs that are a necessity at a certain size).
I’ve speculated that this is the rationale behind some of the recent layoffs. Finally we get some confirmation.
As a heavy OpenAI user, I think it is a premature optimization.
From what I read it sounded like they're making a major shift to investment in A.I. technology, and dropping departments in their company that don't align with that new direction (while aggressively hiring in the A.I. field).
Companies have shifted focus all the time in the past, and axed departments that no longer align with that vision. Meta's doing that right now with enterprise VR as they're dropping that to focus more on A.I. as well (have to jump on that A.I. bandwagon to please them shareholders, after all, which I don't doubt is half the reason Dropbox is doing this as well). But this isn't exclusive to A.I.
There's the free bingo space!
Literally every company is doing the same right now. It seems like the decision was quite simple.
Which areas? Which employment category are they cutting? Hard to tell from the message.
Even though a hundred+ QA products doing this (Chat with your files vibes etc) haven't had much success, they are ignoring the fact that trust is key to document storage, and that we haven't mastered correct context/snippet amalgamation yet - for example, one may want to include both statements that have high cosine similarity, and also statements with clashing natural language inference similarity (contractionary sentences). This gives the context window the statements needed to correctly reason and use logic from different parts of the document. Small things like order (e.g. grouping snippets from different documents together in per-document batches, and also ordering by page number the statements within each document) matter greatly.
What? This is a file storage service.
