cross-posted from: https://piefed.social/c/[email protected]/p/2533091/proton-to-re-interview-staff-about-ai-to-see-who-gets-to-keep-their-job
AI policies are so dumb. You don’t evaluate people’s perfs on how much they use a tool. So what’s the damn end goal of such evaluation??
"Hey! I did nothing useful this year, but all with AI! -You’re promoted!!
I fixed hundreds of bugs and developed 3 new features all cheered by users! -With AI? -Err… not much -You’re fired!!"
The policy my company has implemented is the opposite. AI use is allowed, but no confidential client information is to be sent to any cloud AI service. Logic is that if we don’t have an actual NDA with OpenAI (or whoever), their privacy policy is not robust enough for us to take on that liability.
Interested in your thoughts on that!
no confidential client information is to be sent to any cloud AI service
If the code has been crafted correctly, none should be.
Client info should be 100% data, and remain 100% in production. If you have any dev-side data to test and debug with, it should be 100% procedurally-generated pseudo-data that only imitates client data in terms of gross patterns, and not specific points of data.
Like customer names. Those names should be broadly representative of what is found in the regions the software operates in, in order to demonstrate acceptable data input validation and effective storage compatibility with those names. For example, you don’t want to be storing European names in a DB field that is not UTF-8 compliant, or validating those names by restricting characters to US-ASCII. But any one generated name that just so happens to match an IRL name should be a statistical occurance, and any extended data attached to that entry should (generally) not match any other IRL data for any IRL occurance of that name…
Multiple companies have that approach, I’m not denying there ARE companies doing things right (or at least better).
I work in a company using multiple agents. One of them is supposed to be off Internet, and even with that one we’re not to share sensitive IP.
But Proton is not an isolated case. I bet some execs have invested and promised so much they need to report some good numbers on usage, and you can’t make up productivity improvement if there are people left getting as much done without AI.
For the LLMs killer feature, software development, you don’t really need confidential information 99% of the time.
It continues to annoy me how both haters and glazers will over-simplify the topic by calling everything AI. There’s a giant difference in usefulness, ethics, and quality between LLMs that produce code guided by an experienced developer, and “art” slopped out for greed by some random schmo.
For the LLMs killer feature, software development,
Oof.
I already stopped giving a shit about Proton after the CEO praised Trump, members of Trump’s cabinet, and the Republican party. Then claimed to still be “politically neutral”
Did he actually praise trump directly? If I recall correctly it was his admin pick that was praised
Well the admin pick is even worse lol
There it is. The last straw. My Tuta migration is complete.
Imagine filling this in, in the improvements suggestions spreadsheet. “Who entered this row? I said WHO ENTERED THIS ROW.”
Welcome to Tuta btw. The apps are pretty garbage but it’s home.
;)
Oh and by the by, I rather like the way they tie mail,calendar together in the same app, “tabbed” style, sort of. Give me a year and I might have some complaints, but for right now my only complaint so far is it seems the “Revolutionary” level subscription isn’t “high-falutin’” enough to be allowed to import old mail. Just prompts me to upgrade my account.
Their youtube channel has a year old video about importing that says “all paid tiers in the future” but I guess we’re not at that future yet.
I believe “one click import” or something similar was a recently new feature, or something like that? Like importing, or moving from, other providers. I might misremember though.
Yes sadly it appears that’s the feature that just opens a prompt to upgrade my sub. I’m fine with a local archive of my other account, so for me that’s fine, for others it might not be. Thank you though!
Ah, the “one click” thing was for exporting from Tuta rather than importing, my mistake.
Either way, I found this:
We’re thrilled to be able to offer email import (in beta) currently to Legend and Unlimited users. In the future, email import will become available for all new paid plans but as we have just released it in test mode, it’s temporarily necessary to limit the number of users importing.
What worries me a little bit is the “will become available for all [NEW] paid plans” part, as if old paid accounts won’t be getting this new feature. But I hope that’s just a phrasing mistake/interpretation error.
I’m currently on Legend because they had some deal a while ago where you could get on that plan for the same price as Revolutionary, but I don’t see import in the mobile app. I could check the desktop app later!
Edit: yeah, when I’m using the desktop app, I see the import option, and clicking it does not prompt me to upgrade my subscription. (It just shows me the screen to import email as you’d expect.) Again, this is on Legendary for the time being. We’ll see what happens in March when I have to downgrade to Revolutionary, if clicking Import prompts me to upgrade. I hope not.
The AI stuff is so frustrating. I’m terrified to really talk about how I feel about it, because my opinions are too nuanced to be shared in a bullet point in a performance review. Everywhere I’ve been lately, it’s been about how much you use something, not about your output.
What is the computer equivalent of “walk fast and always carrying something”
This looks too l33t. I fear if I had this going, typing like a maniac, everyone in the office would gather around and start reading the code.
“Everyone look busy”
There’s a reason we keep the companies services separated, so at least if they do manage to generate enough moral outrage in me to move I only have to migrate a dozen email accounts. Never touched the rest of their service
Lol wtf proton was supposed to be good and I was moving to them, ffs, smfh!
If you read the Proton response to this article, it actually does seem like trying to figure out what’s going on with business process in a bog picture way. Senior managers are interviewing, too. Not all staff are interviewing.
Personally, I think it’s just as much about seeing who can’t QC the slop out off their code as anything else. Maybe. Or maybe I’m wrong, I dunno.
How about don’t use AI to begin with so you won’t have to weed anyone out due to “potential slop”.
If you haven’t used it professionally, LLMs can and do provide huge help. Making this a specially all or nothing argument only highlights a knowledge and experience gap on the part of someone making that argument.
Buuuuut, only an idiot asks for anything better than a starting point or a check of existing work. People relying on it to produce whole cloth work is for hobbyists and sloppers that don’t care or know any better.
You ask why they need to use LLMs at all, and they host their own, Lumo. What clear, financially solid argument can you make that LLMs are an objective net negative to every company? If they see benefits from a headstart on coding and ability to test their security, why demand they take that away when no human code is 100% perfect anyway?
I haven’t used it professionally, but I still know they provide huge help. It’s no secret. But my statement was simply following the logic of the previous commenter, who was saying that Proton are trying to weed out the people who aren’t able to QC out the slop. But if you are in need of QC to remove the slop, you are probably overdosing on agentic coding to begin with.
Anyway, “only an idiot asks for anything better than a starting point”, sure, definitely agree. But what percentage of being a professional developer at a company involves prototyping something new, and doing “starting points”? Most development is just tacking on new shit to old shit. Clamping shit together with other shit. I guess that’s the issue?, which causes people to still use AI because they are pretty much forced to, sometimes literally, and so actually do ask for more than just a starting point. And then here we are…
I’m coming back to this only because a relevant post came up elsewhere on Lemmy that’s maybe worth both of us seeing in the context of our own opinions.
I fear you may misunderstand their goal here- direct quote:
“We need people who can translate AI’s capabilities into better products, stronger execution, and new ways of working — and who approach this change with curiosity, judgment, and urgency,” the email reads.
They specifically want to “weed out” the ones who ARE NOT using AI
Welp, even worse. Fuck em.
NOW we agree. Fuck em hard.
I for one am not sticking around to find out.
Imagine still carrying water for a corporation.
Aren’t they a foundation?
No, they have a for profit company and a nonprofit entity that is a major shareholder. They operated just like a corporation with all the typical warts.
Yuck, they really make it seem like they are just a non profit.
They are definitely gross, but so is every corporation. The real lesson is stop giving companies money.
Imagine being so disillusioned with the world that you assume that no one can have an opinion and think critically about two opposing views without it necessarily being some corporate brand fanboy bullshit.
Here, let me help you. Proton doesn’t give a fuck about you.
Yeah, was that supposed to be profound or something?
As the 1% that actually pays, I’m invested in them succeeding and not enshittifying so I don’t have to take the time to move my email hosting. End of story. I have my own email domain, so I can kick them to the curb any day I like. And expect to have to one day.
Did you need to complain about it some more? Does baby need a Binky? Or did you only ever use the free thing and it didn’t meet your expectations of what free things should be? Sorry your free thing wasn’t spectacular enough :(
Knew it!
Knew what? That someone doesn’t want the thing they pay for to get enshittified? That it’s smart to make plans to not let any company control how you communicate?
Wow. Such insight.
That you were a paying customer hence your rabid defense of a garbage company. Every time I get massive pushback it is always someone that has bought into the security theatre corporate solution that they sell.
I am sorry you continue to give them money. Take care!
Can someone just start a damn privacy-first email coop already?
Runbox is a Worker coop based in norway
Do you have more info on how the company is structured? Everything I can find on their site just says it’s “employee-owned”, but I don’t know what that means.
IMO, using AI to advise or guide you is the way to go, as it’s not actually touching the code.
But AI doing the work is a disaster that will lead to epic amounts of technical debt because it is impossible for AI to utilize judgement or taste.
I have been using AI more and more lately to just advise, to find, to make me aware of other options. I still do 100% of the work. So while AI hasn’t made me faster, it has made me more flexible by kicking me out of ruts and making me aware of alternative paths to reaching an objective.
I will not let AI touch my code until/unless it acquires judgement and taste. But until then, it’s turning out to be one hell of a gopher and rubber ducky.
What you’re saying is you’re happy to pay a subscription fee to use an electricity burning earth destroyer run by an evil tech oligarch rather than put a 59 cent rubber ducky on your desk to reason things through with?
May the flying spaghetti monster have mercy on your soulless husk.
Why should I pay a subscription fee? The AI is not touching my code, not generating any of it. I use the free versions to ask questions or give it snippets as examples, and go from there.
Sure, I’m beginning to bump into Claude’s free ceiling, but there are so many other options out there that I just partition my questions to the most appropriate model for that question.
There may come a time when I actually sign up for a paid tier, but it hasn’t happened yet. Until then, I will continue to exploit the parasites that fund this largesse.
And honestly, I won’t be upset if it all crashes and burns – it’s only a tool for me, not a crutch like with so many other devs as of late. Almost all of them will have regressed all the way back to junior-level skills if they had to hand-code all over again. It’s actually kind of sad to see such skills erosion and dumbing down that’s happening IRL.
Every technology you touch hurts the environment Could say the same about electricity
The device you are using right now to write here is with 99℅ certainty created by exploitation of people - why is it OK for you to usw that devicd but not for other people to use ai ?
Every technology you touch hurts the environment
All life causes pain and suffering to other life.
Therefore, the logical end-stage conclusion is to trigger a false vacuum decay event that erases all of existence throughout the universe and so relieves all suffering.
Until then, I will continue to exploit the parasites that fund this largesse.
Did you miss the part where you’re also exploiting all the people hurt and the environment polluted in the process of building and operating power lines, power plants and data centers?
Or the part where the crash would come down on the neck of taxpayers? The parasites will be fine when the dust settles, chip and cloud resource vendors will have made a pretty penny, the executives’ salaries will have built a wonderful cushion to land on. But if people lose their savings or pension funds need bailing out, that’s a textbook example of the “socialise the losses” part of capitalism.
Or that they might actually get away with racking up the prices far enough to become profitable and stabilise the bubble, which would most certainly also come at terrible cost to the individual worker, whether or not they personally used or endorsed these tools?
Nothing about this should be cheered on, neither the possibility of the bubble stabilising the prospect of its crash. The only winners in this farce of a gold rush will be the ones selling the tools for everyone else’s misfortune.
The sooner it pops, the lesser the damage, but damage it has done and will do.
Use it, if you must, and hope that you won’t find your own skills having rotted away after all, but don’t delude yourself that it’s some noble act of sticking it to the rich.
I agree. If I use AI is isn’t to write a report. I write the report and then have AI review it to see if it catches anything I might have missed like repititions or something, and make suggestions like make something more formal or scientific or something.
If I need an excel formula I will make it and if it doesn’t work exactly as I want it to I will ask AI to give feedback on the formula and see why it might not work.
That is increasingly concerning, i’ll start exploring other options. Can’t believe they’re rotting from the inside.
I use Lumo all the time as it’s the only LLM that respects privacy right now. It’s not a top model but still has web search and is very reasonable for small private issues. From Proton replies seems like this is overblown and they’re looking for internal people to work more on AI stuff.
So if I’m reading this right, Proton is re-interviewing their entire staff, focusing on how they’ve integrated AI into their workflow, and if they aren’t using AI as much as management wants they’re likely to lose their jobs?
That is extremely disappointing.
Posteo is suddenly looking very shiny.
Give it 5 years and that shiny will ware off. That’s the damn problem. Nothing last forever and everything goes to shit eventually.
At some point you just gotta settle for the devil you know.
If you still have the energy to most posteo is fantastic from what I have heard tho.
I’m just tired man, I just want to stop having to think about politics in my fucking email.
Does posteo support custom domains yet?
No and they don’t plan on it, as far as I know.
Ok yeah, and I understand why that is. I just need a custom domain or I’d be right there.
If all you need is solid email hosting, I can vouch for Migadu. I’ve been using them for years with my own domains. Been perfectly reliable at reasonable costs.
Tuta is good. Waiting on their drive.
Tuta cancelled my free account because their automated system detected fraud. This occurred after 24 hours of updating my contacts within Tuta and signing up for some services. I appealed and they only apologized for the inconvenience but kept the account closed.
Other people have posted similar stories but after a YEAR of use their email randomly because inaccessible. So Tuta isn’t a viable alternative for me as a Proton user.
Damn, second negative response. I guess I am leaving them but I just do t want proton - they give me the ick of wannabe big tech.
I just want a drive, calendar and email that is private.
Posteo seems like the only good option there… Sadly they don’t allow custom domains because of some hand-wavy “privacy” crap that boils down mostly to vendor lock-in like tuta.
There is a new belgian co-op company Nubo trying to break into the market, but they are tiny and running a nextcloud instance on-prem for storage. Performance is likely to not be great.
vendor lock-in like tuta.
Higher tiers allow custom-domains.
No IMAP, doesn’t support open standards, proprietary encryption, no compatability with 3rd party apps/mail clients. You get forced to use their app and have none of your mails locally. Also, since their encryption is proprietary, they only support password less encryption between tuta users, so 99% of users will only ever have encrypted mail with the tuta advertisements/announcement (though they market the “encrypted” angle extremely hard.
Not to mention all of their apps are exclusively web and electron and perform badly with a very very barebones user interface and options. Also the search is pretty bad. All things that would be fixed if they support open standards. I really dislike using their apps and interface with my account.
Apparently Tuta sucks, which really sucks as I was halfway through migrating my email to them. Now I have to start over with Posteo and wait for them to do something stupid
TIL about tuta, aside from my shitty experience with their mobile app now.
This came up in my financial news feed. If you’re a publicly traded company you want to be a growth company and not a mature company. The latter group are stable and provide steady dividends. The former have shares that become more valuable and increase your worth.
This is how we get people pivoting from sparkling water to AI.
And it’s how we get software companies that have a solid product suddenly deciding to ruin it by adding AI.
from sparkling water to AI
If you’ve got to prep for a job interview either way, then you might as well do a few more.
Well looks like I may need a new email and password manager soon.
