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.
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 really don't think there are many existing companies that "pivoted" to blockchain and NFT. Certainly not a majority.
Blockchain there were a lot of hiring and experiments for sure. Actually to try to pivot a company to it after these experiments, very little.
I have seen many companies announce that they would look into blockchain and NFT. Which is exactly what Dropbox is saying now for AI.
https://about.fb.com/news/2022/05/introducing-digital-collec...
https://www.washingtonpost.com/video-games/2023/01/04/square...
https://decrypt.co/121872/ubisoft-rabbids-nfts-sandbox
https://vanguard-x.com/blockchain/nfts-are-coming-to-youtube...
Lacks an exclusion mechanism, so icons leak in
Includes STL files as photos and fails to display them
Sometimes misses photos for no obvious reason
You guys are reading way, way too much into this press release.
People here are acting they've abandoned their existing product and is betting the company on AI, when it could just as easily just be what every single major company has said recently ("we're exploring AI").
LLMs do. Guess we'll see if it's really flexible enough to fit into a lot of business models, but it's not exactly the same thing.
I don't see it as the wildest thing that say, a corp using cloud storage might want an LLM trained on all their documents so instead of searching for where X employee who left the company 3 years ago stored some information that was on page 9 of a document from 2017, you just ask the model and it pulls the info and links you to the source doc.
Well Blockchain promised a lot. I remember that in the beginning, I found it interesting and thought it would be the next big thing. Now we have had more than a decade to realize that Blockchain doesn't solve anything other than cryptocurrencies (and it's not clear if we want cryptocurrencies).
So yeah, I could totally imagine that many promises of LLMs today are similar to those of Blockchain back then. Let's see how it evolves in the next few years, I guess.
Then again, never try to convince someone of something when their salary depends on believing otherwise.
Uniswap likewise has billions in daily volume. Again, that's significant. I really don't understand how the hacker news groupthink can see a piece of technology that enables something completely brand new - decentralized market making - getting used daily to move billions of dollars around and say "there's absolutely nothing there".
Maybe you think there's no killer app because you hang out on upvote-centric site like hacker news and reddit where people downvote the hell out of things that contradict their narrative. Hacker news is a groupthink bubble.
There's many other examples. Look at how often people brought up energy usage and GPU shortages when NFTs get brought up. Then look how often it gets brought up when LLMs get brought up. It's night and day. None of these people cared about the environment. They didn't like crypto so it was a talking point against it - that's it. Because now, ETH is on proof of stake but everyone and their dog are buying 4090s so they can make waifus with Stable Diffusion. And all those "you're melting the planet people!" are conspicuously quiet.
What exactly is the killer app of LLM? A bunch of writing tools for SEO spam? Does anyone seriously read what ChatGPT writes and think this thing could do real literature or journalism? Everybody wants to use AI to write, nobody wants to read what AI writes. The only real "killer app" I'll give it is Github copilot. Most everything else is froth.
When NFTs were trendy, I read constantly on here they were beanie babies, that they were naming a star, whatever. You know what bubble I saw excited about them? Not the tech bros, but artists interested in a new way of monetizing their work. Go read Grammy award winnner RAC's twitter - he addresses people saying he "scammed" people by @ mentioning those he sold NFTs to and asking if they feel scammed. Fans of his bought the NFTs so they had a digital collectible representing his album, he got money to make art.
When Babe Ruth signs a baseball, it's still just a dingy baseball, but it has emotional significance to people. NFTs were the digital equivalent of that. I'm not surprised techies that have no art in their lives struggle to wrap their head around that.
But what is disappointing is that now Stable Diffusion takes a bunch of art that real humans tirelessly made, uses some neural net to rejigger it, and tech bros will die in the trenches making sure we legitimaze this quasi-plagriasm. For all the talks of NFTs being scams and generative AI being substance, I see one technology that incentivized real humans to make real art and another technology that does the opposite, takes money out of artists pockets so people can use algorithms to lift their style.
I'm not trying to oversimplify these topics but it gets tiresome reading the exact same tired talking points about generative AI being substantive and web3 being all scams when there's strong arguments indicating otherwise. It's just impossible to read them when on sites like these they get faded out because you're not allowed to have divergent opinions without a bunch of defensive dorks downvoting you into oblivion.
There's one last important point I want to make. Sites like Reddit and Hacker News and especially mainstream media are unduly influenced by big corporations. True blockchain use cases give power back to the individual. Sure, there's plenty of VC pump and dump shitcoins in crypto such as Solana and NEAR but the long term real utility is about the individual. Meanwhile, LLMs are almost exclusively trained and served by massive corporations, the same massive corporations that can influence the media you read. If a journalist wants to write an article about crypto energy usage, they'll write it. If they want to write about LLM energy usage, you have some of the richest, most powerful people in the world with some of the premier PR firms in the world who can influence that journalist and that publication in all sorts of ways. So critical stories about AI are more likely to get buried than critical stories about crypto. Keep that in mind.
To me, the biggest difference between the AI hype and the crypto hype is that crypto had the skeptics and the critics in the room - as it should. But AI is badly, badly lacking those skeptics and critics.
And the idea that an NFT has any "emotional resonance" is just weird.
> I really don't understand how the hacker news groupthink can [...] say "there's absolutely nothing there".
Blockchain doesn't solve anything other than cryptocurrencies that was not solved better before. But that leaves this one thing: cryptocurrencies; blockchains does enable cryptocurrencies in a way that was not possible before.
The problem I have with cryptocurrencies is that I don't want them. That's not a technical argument, that's more a preference of what I believe society should do.
So there is not nothing, but web3 is bullshit IMHO.
> You know what bubble I saw excited about them? Not the tech bros, but artists interested in a new way of monetizing their work.
Well, it's hard not to be excited by something that you don't really understand but that may bring you money. Doesn't mean NFTs are desirable. BTW you talk about them in the past, so somehow you do realize that apparently they were not desirable enough to survive, right?
> Meanwhile, LLMs are almost exclusively trained and served by massive corporations
I totally agree with that. I don't want Big Data to steal my data, use it to train models, and sell that back to me. I wish there would be a way to account for copyright and licensing in a decent manner, but I fear that the rich will win (that's capitalism, right?).
Just as much as I believe that Bitcoin made more harm than good, I believe that generative AIs have the potential to make more harm than good. The problem is that companies don't think about whether they should do something - only whether it is profitable. And people typically love to not think about it and just happily try and support all those cool techs.
My non techie wife and all her friends love ChatGPT and use it daily. My kids friends use it to do all their homework. Which is pissing off my kids as I won’t let them use it.
Not nearly everything in AI is ChatGPT, not quite all is NLP for that matter, there's still plenty of interesting things which are improving. Dropbox might well be pivoting to improve their e.g. photo/video offering with AI, but they might as well have started that 5 years ago.
The current “wave” of AI and deep learning and whatnot has been going on for almost a decade, longer than blockchain and definitely longer than NFTs if you ask me. Not saying that it’s not overhyped, but it’s definitely not a new hype.
For Dropbox, I predict that this will go exactly the way Dropbox Paper went.
I also had a batch of photos to process from an event and updated my Lightroom. The new Denoise did a really nice job with these relatively noisy/highish ISO pics. And the Adaptive preset to punch up the subject (speakers at an event) was pretty nice too.
Not game changing IMO but a nice workflow addition.
NFTs on the other hand always felt like beanie babies. Blockchain in the enterprise distributed ledger sense felt like it had some potential use cases although I was always pretty skeptical which seems to have been justified.
I do agree with this observation of yours, fwiw.
If I succeed, do you want an invite?
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?
In that spirit, here's me doing this about AI at Dropbox:
- Get consent from users and train a personalized AI on user files, to answer questions like, "Where is my meeting with Thiago?"
- Provide built-in support for number crunching on unstructured meeting notes: "How much revenue did we make, from last quarter's notes?"
- Automatically fine-tune a model using your own writing so that ChatGPT talks more like you
- Apply an AI to recognize security threats in your files("Hey you might have stored a secret in plaintext in file Y. Do you want to encrypt it?")
* find someone that will actually want to pay extra for their subscription on that.
What if it’s more about using customer data to Dropbox’s advantage? Combined with the telemetry you can’t turn off, they can mine the shit out of users.
I can see DropBox making some interesting features for their products around AI. That’s going to require manpower to do it. Manpower that’s probably a little nervous about staying at a company that just did layoffs.
This is genuinely blind to anything that's happened in AI over the last 6-12 months. The data that is the most valuable to AI is exactly all of the random, unstructured data (product requirements, decisioning documents, engineering designs, process documentation, etc etc) that has been shared across teams.
> I just can’t even come up with plausible uses for AI that meaningfully change the company enough to require a layoff.
requiring a layoff no, certainly me neither. There is definitely a bit of "compliment sandwich" happening in this letter. The layoffs were likely just a long time coming for a company that has been rather stagnant for quit some time.
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.
In an enterprise, it can be extremely difficult to ensure your permissions are what you want them to be: that people can share things easily with the right groups, but also that sensitive data is not inadvertently exposed. Dropbox in particular excels at sharing documents with others outside your company, but that is also where there is obviously the most risk.
Currently, I find Dropbox's enterprise permissions management tools pretty difficult to use. There are loads of options and it's too easy to get something "wrong" if you inadvertently miss checking the right checkbox. It's not hard for me to see how AI tools could help improve this situation, and especially to provide additional capabilities in the DLP (data loss prevention) space that would make it easier to detect misconfigured access.
This seems the exact opposite of what the current crop of AI is good at.
Yeah, last thing I want when trying to share a file with people outside of my organization is some opaque kafkaesque chatgpt-wannabe model telling me "I’m sorry Dave, I’m afraid I can’t do that".
1. Letting users enter desired permissions setting in a natural language, and then the AI recommending checkbox settings, and, importantly, explaining these settings.
2. Useful as a monitoring/alerting system for DLP. Most DLP systems already use some sort of machine learning for identifying sensitive docs.
3. Easily running "test scenarios" to show to admins who can and can't get access to particular docs.
There is a huge chasm between "AI owns all my permission settings" to "AI can make it easier and more robust for me to understand what my permission settings should be."
> Currently, I find Dropbox's enterprise permissions management tools pretty difficult to use.
These two sentences were one after the other. So which way is it?
That’s only because it failed (and the jury is still out on that). If it works they’ll conquer an entirely new market right from the start.
Certainly a much better long term strategy than hoping they can keep up their advertising revenue.
The "metaverse" is ironically set up to be Facebook's walled garden. It has no other use, it cannot be repurposed or made modular in any way. It's not a standalone product that say, integrates with Google Ads, something that they could try to offload to some other company.
It's Zuck's baby, and in the middle of slow and protracted crib death.
The last 15 years of tech says otherwise. Tesla is the most obvious counterpoint to this example. I don't think the Metaverse is useful, but my issue with Facebook isn't that he poured 30B into what was essentially R&D.
Facebook has their own platform but it's built on top of others so it can be choked off. Facebook wants its own OS.
I would have thought HN would celebrate a company that's publicly decided to stop fucking with the Wi-Fi.
If I built a company and eventually found myself in extremely high value (ie, 9 figures), I'd be looking to retire ASAP. That would be it. I won capitalism. Time to spend the rest of my life on Maui sipping Mai-Tais on the beach.
Why anybody would want to continue to work when they already have enough wealth to live an extremely lavish lifestyle without working ever again is beyond me.
With that perspective, the best thing to do is keep on pushing. Despite whatever pain and headache that might bring, it's one act our short lives can use to push against entropy and the dying universe.
If I ever come into such wealth, I'll spend it all solving as much as I can through my direct involvement.
If I ever came into such wealth, I think I'd spend my time going back to college to study linguistics. Maybe document some dying languages. Try to contribute to humanity a bit.
Pet hypothesis: A lot of us say that, not being in the position to. But when you finally exit the working class and are no longer required to work, people tend to work on passion projects instead. Sitting on the beach gets boring after a while.
Is probably extremely boring - I have never had more fun than when solving problems, life without working on something interesting sounds insufferable to me
How long would I be able to sit on Maui sipping Mai-Tais? Getting massages, ... and more ... whatever I want, whenever I want it ... that is a curse actually.
Surrounded by fake people everyone wanting a piece of the wealth trickle down to them.
On a personal experience, I was a lot happier when I bought my first house, a rundown estate that was the cheapest on the street that then I had to work to improve upon it, ended up replacing most inside and out, some parts I learned to do myself - I ended up remodeling it piecemeal but so, so satisfying
My second house is very respectable, newer, larger, technically incomparably better, it is not even the same ballpark ... yet my new house never provided me with the same joy.
Be careful what you wish for.
I'm fairly sure that there are many people like this out there, but for others that just... isn't a problem?
Personally, I quit my last job because I felt like I need to learn some new technology, work on a few personal projects, write blog posts and so on. Since, I've done a little of the above (in addition to improving the security of my homelab and running my own CA), but also took a month off to just enjoy watching videos, reading articles, playing video games, as well as do physical labor around the farm.
And frankly, everything feels enjoyable so far - both intellectually stimulating tasks, physically intensive labor like cutting trees and chopping firewood, as well as passively consuming content, or even zoning out a bit occasionally and having a slow day. Though I can't comment on how long each of those would be satisfying for, that also depends on the person.
I can easily imagine someone enjoying a lavish and relaxed lifestyle and never getting tired of it, though.
There is strong correlation between people who find themselves in 9 figures (outside of inheritance) and people who have completely opposite worldview from yours. If you make money to retire to Maui, you will take the first off ramp which allows you to do it which is way before you get into 9 figures.
Some people might like to use their money to take over an institution like the World Health Organization to implement their eugenics program that they dreamed about as a little kid.
Others might be interested in buying up all of the farmland so that they can control the food supply.
You just never know how you will respond when you acquire that kind of money.
That's just your hindsight bias. Lot of companies have spent billions of dollars chasing apparently worthless product ideas. Some of them turned out to be big hits - electric car in 2009, social media based on 140 characters posting limit, online bookstore, a website to check out pics of your classmates and poke them etc.
I like the fact that Facebook was audacious enough to spend 40B dollars on trying out something new. If successful, it would have opened up a huge new field for tinkering, just like the wave of social media companies in 2009-10.
Ooh betting on a failure? That is the easiest bet ever. If I bet on everything failing, I would win 99% of the time.
Literally every company. That doesn't mean that a strategy doesn't exist and is actively curated and maintained by people whose full-time job it is to do so. It's always an educated guess at best. But it's pretty silly to come into a thread like this and be like "psh they store files what could they possibly do with AI".
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.
Maybe DBX sucks at executing theirs. Maybe their strategy sucks. I don't know. I'm not claiming to. All I am claiming is that there is some strategy at the leadership level right now at DBX that appears to involve AI - and any comments beyond that are purely speculative.
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.
Meh. You could've said the same thing about companies leveraging dot-com, or going mobile first. All these were a net gain for investors and society.
"The future is now old man."
You'd be surprised.
>"The future is now old man."
Well, September 4, 476 AD was also "the future" of much better pasts for Romans.
Something else?
1. that companies should have specific product areas and focus, and not pivot all over the place, as mere profit-seeking machines.
2. that companies embrace all kinds of BS trends, hyped technologies, and empty vacuum BS solutions looking for a problem (from NFT, web3, metaverse, blockchain)
3. that companies (as in this case) also embrace valid technologies (like LLMs) that have no real application to their products, and will just retrofit them in some ugly BS way nobody asked for, just to send the signal "we're still cool, we're getting on the next bandwaggon, buy our stock"
4. that the current behavior (actually, for decades now) is not about "sustainable growth and healthy profit" but about stock market games, and about endless unsustainable growth that destroys their core offerings and alienates customers
2. So what? People embrace all kinds of trends. And companies are made of people. This is basically a tautology...it's only a trend of people are embracing it. Also zero of your examples are the topic here.
3. This goes back to my original post...we don't know how they plan to use LLMs. It might be useful, it might not be.
4. Just pure cynicism. Some of what you're saying happens at some companies. But not most.
Yes, and I'm saying they shouldn't.
>If they fail, they will fail. Companies are mere profit seeking machines. Perhaps you are thinking of non-profit orgs?
No, I'm thinking of the idea of the civic responsibility of companies, and of a market that serves society, not just maximizes profit. Perhaps the barbaric 2023 is not ready enough for it.
>So what? People embrace all kinds of trends. And companies are made of people.
Yes, including BS trends, detrimental to them, their families, or their communities. And, like companies, should embrace less BS trends and more good trends.
>This goes back to my original post...we don't know how they plan to use LLMs. It might be useful, it might not be.
We do know it has nothing to what people got into Dropbox for: simple file sync and offline copies.
>Just pure cynicism. Some of what you're saying happens at some companies. But not most.
Simple starry-eyed market fanboyism. See how two can play the ad hominem game?
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.
My message is that OP doesn't know DBX's product strategy. I don't either. I never claimed to.
> Then you go on saying this is a common strategy and is well suited?
Sorry, where did I do that?
> Do you understand the strategy or is your comment as worthless as you suggest in your first paragraph?
No, I do not. I never claimed to.
> 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?
I have no idea what their strategy is. Neither do you. Neither does OP. That is all i said.
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!
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 not only does your comment not contribute to any interesting conversation, it can also likely be wrong as well. Some times when product decisions seem questionable, they really are questionable.
This is just some rando internet person making reductive statements about a company whose current offerings they are unaware of and future product roadmap they have never seen.
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.
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.
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.
>I remain unconvinced of the usefulness of the Syncthing sync model on iOS
https://forum.syncthing.net/t/syncthing-for-ios/16045/4
So basically it's not an alternative at all to Dropbox
And to be honest, most of the files I sync, I never access on a phone anyway. So it would still be useful even without Android support.
Not sure how common my setup is, But I have desktop both at home and at work. And I use git and synching, to sync most of the stuff seamlessly. That said I mostly work from home nowadays, so I mostly use my work comp to speed up compilation, or run tests, dockers, databases, etc.
* Not trying to be an asshole. There are plenty of iPhone only app, that I sometimes** wish I could use, so I know how you feel.
** But not too much, otherwise I would switch by now :))
With end-to-end encryption you can transfer data with an untrusted server as intermediary, which syncthing doesn't support.
(I don't have proper backups beyond Dropbox, because I'm pretty happy for stuff I've actually deleted to stay deleted. Just, in case of a hard drive failure or something similar, I can sync back from Dropbox. That's all I feel I need. Maybe there's something better?)
An FTP server already wouldn't serve my needs given that there's no such thing as mobile sync to an FTP server.
The irritating thing with a lot of software is that they start out like a hammer and in 10 years it's a multi-dimensional VR rangefinder with email. Well that's super cool but I want my hammer back.
The time macOS changed my default to having all my files in the cloud rather than local was a truely scarring experience.
The price is high and ‘family sharing’ is a minefield.
You’d probably be better off using a different service if it bothers you so much.
“I used to be ‘the target customer’, but then they changed what ‘the target customer’ was. Now I’m not ‘the target customer’ anymore and ‘the target customer’ seems weird and scary. It’ll happen to you!
Isn't it in the nature of tech industry though? If a company "just is", it doesn't need to do any R&D work, so it's not a tech company any more.
Rule 1: Be an enterprise account that uses the "Call us for a quote" option, and does not complain about feature changes or removals
Rule 2: Don't break Rule 1
In other words, encourage this mindset wherever possible, it's going to be what causes the next wave and spike in demand for SWEs.
ChatGPT can come up with completely made-up answers that just seem right.
I describe AI as a new intern or assistant. Pretty useful, but don't bet your company on their answer.
This leads back to one of the AI promises if the earlier Internet, before WWW when Usenet, Gopher, and a number of other ideas, were where we were heading, and for a time after WWW's popularity before search engines hit certain quality/coverage levels: an intelligent search agent trained on information about you that would go off and scour the various data resources for you and return an optimal set of results after a while, perhaps even idly looking around and finding things you might search for but hadn't thought to ask it for yet.
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.
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.
Can a utility just stay a utility?
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.
Only companies where owner is top boss can, sometimes. The moment boss has to answer to investors, growth is required.
Capitalism - regulation => utilities's entire point is to capture a monopoly to then abuse it.
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.
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!
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!
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.
Dropbox also never hired at a rate beyond their rate of revenue growth.
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).
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.
I guess they have Paper, but last I used it, it felt like abandonware.
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."
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 can think of plenty. Text-based image search, "find the latest version of 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.
combined with the biden depression, this makes it a good idea to fire employees and gain this new source of income.
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?
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.