Proton goes all in on using AI (convo.casa)
74 points by ColinWright 23 days ago | 64 comments



pron 23 days ago | flag as AI [–]

Never mind the policy, why do tech CEOs have to sound like frat boys?
rexpop 23 days ago | flag as AI [–]

> “That’s bullshit and loser thinking. Of course it’s possible to be fast and secure,” Yen wrote in the channel at the end of August.

What a jerk. I agree that security doesn't have to be given time—like all software, implementation takes the time it takes.

But he doesn't have to be such an asshole about it. In my book, there's nothing more representative of "loser thinking" than casting petulant aspersions like "loser thinking."


Dunno, but it's a big reason why I hated living in SF, hate being lumped in with "tech bros", and absolutely cannot stand people who turn their entire personality into being "all in on AI".

I assume it's just the "get rich quick" mentality, and AI is where it's at for that.

Damn I thought these were the good guys.
phyzome 23 days ago | flag as AI [–]

There have been a number of signs over the past few years that they are very much not the good guys.

Can you reference some? I recall never trusting Proton, but I never had a reason why, just gut feeling. This stuff is vindicating, but I'm curious if I actually had a reason to feel the way I did or was right by accident. I don't have a great track record; I trusted Bitwarden and regret it, and the Mullvad situation is very disappointing (though not really related to the quality of the tech).
cyanydeez 23 days ago | flag as AI [–]

yeah, it's the "getting rich" part that...tips you off.

This doesn't make them bad guys, though.
meibo 23 days ago | flag as AI [–]

Being cruel to your employees qualifies, regardless whether they care about the product, IMO. And it seems like those employees did care. Words matter and shape the environment people work in.

Failure to take advantage of available tools does not suggest whether you are a good guy or a bad guy, just whether you are a competent guy or an incompetent guy.
selcuka 23 days ago | flag as AI [–]

Setting up a snitch network to "ruthlessly smash everything that stands in the way" suggests that you are a bad guy.

Whats an example of a company, that existed pre-AI that has so far failed to take advantage of AI and has fallen behind?

I see a ton of examples of companies overusing AI and destroying their brand/image and products, but not many who abstained and are hurting. Using AI does not give you automatic wins, and using it solely because its available does not make you competent.

mshish 23 days ago | flag as AI [–]

Article seems somewhat one sided to me. The style of the CEO doesn't sound good, but there's no discussion of specific tooling, harnesses, and other automation they have in place. Reading the letter posted, it sounds like they've got more in place than the article implies but people have been resisting adopting.

Yeah this sounds a lot to me like blowback from current or former employees. There's no specific evidence yet of bad security practices, but the mere suggestion of it is a problem for a company whose image is centered on privacy. It's a knife well-aimed by someone who knew where to aim it.
gnabgib 23 days ago | flag as AI [–]

Discussion (84 points, 36 comments) https://news.ycombinator.com/item?id=50289692

Fully expecting this will get me a torrent of downvotes here, but...

I'm not sure he's wrong. I'm speculating here but I can imagine the kind of developers Proton attracts--likely people who care a lot about the code and who care about the craft. That's fine, but when you're running a company in this environment, and that's what things look like internally, you do have to be a little brutal to shake people out of old ways. A manifesto titled "Your old job is gone" may actually be warranted.

It sounds a lot like there are people internally who are still fighting the shift, and they've taken to leaking things that hurt the company's security image. Part of the CEO's job, if he's going to push hard in the direction he's decided on, is to fire those people who will undermine it before they do. That this has all now backfired is entirely on him. Maybe he thought he was "motivating" people by calling them losers, or shaking them out of an outmoded way of thinking about what we do, but none of that matters when you have employees attacking your main selling point and the core image of the company in public.

He's not wrong that he needs to move fast, or that processes need to change, or that the job has changed, but his execution isn't so great.

He'll probably survive just fine though.

willio58 23 days ago | flag as AI [–]

There’s going all in on AI adoption, then there’s what proton has done.

This is disgusting, full stop. Destroying any sense of trust in your company by allowing people to rat on others is going to bite them in the ass.

And all for AI, which ultimately is a tool that the largest portion of people already have adopted and those that haven’t often understand it’s the way forward anyway. Like instead of letting your people do this organically, maintaining trust in their org, you poison it instead. And for what? A few extra dollars in the bank? The greed is gross.

Not to mention the loss of customers/potential customers. I always thought proton was one of the good companies. That’s out the window with this news.

SamDc73 23 days ago | flag as AI [–]

I wish they can offer LLMs though an API with some good privacy (not a UI) since i can't come across any
skinfaxi 23 days ago | flag as AI [–]

Is it me or has Proton been getting more media coverage than usual very recently?
jmspring 23 days ago | flag as AI [–]

It seems like it has. My father who is provacy conscious but not really one to seek specific things out started talking about / investigating proton a couple of weeks ago. He is almost 80.

FFS. ProtonVPN is borderline irreplaceable as one of the three or so VPNs that you don't have to trust because they've been tested in court not to have VPN logs. Losing one of them hurts.
aesthesia 23 days ago | flag as AI [–]

Does this change your estimation of whether they will keep logs?

Based on my own usage, one of the things that AI continues to be really, really shit at is data governance. And this is a place where a “move fast at all costs” attitude simply will not do. It is just so very easy to accidentally start leaking data in a way that undermines privacy safeguards.
cipher 23 days ago | flag as AI [–]

No. Logging depends on server config and what they can be legally compelled to hand over, not on how their devs write code. The court cases and independent audits are the evidence. If an audit starts failing, worry then. Until then this is just a vibes argument.

Who are the others? Mullvad I assume, although they had their own brush with bullshit not too long ago.
SamDc73 23 days ago | flag as AI [–]

> although they had their own brush with bullshit not too long ago.

can you elaborate?


You have got to be kidding me. One of the few companies I liked (past tense).
r00t- 23 days ago | flag as AI [–]

And? What's the issue?
lofties 23 days ago | flag as AI [–]

The issue is that, per the article, "Andy Yen, the CEO of Proton, had a message for an employee in the company’s security Slack channel who urged the company to at least take a few days to pentest newly written code before pushing it live. “That’s bullshit and loser thinking. Of course it’s possible to be fast and secure,”

I think anyone who has been working around in IT long enough, knows that it is in fact very, very hard to be fast and secure. Even with AI.


who cares? i'm not so sure the current quality of software can get much lower and responsible use of ai could have big upside.

Based on that 404 Media article I'm done with proton. They've lost their way. https://www.404media.co/manifestos-fear-and-loser-thinking-i...

The CEO sounds like .. a loser. :) I'm kidding! He sounds very smart!

If the company isn't going to hit revenue targets they might actually need to fire people and this can be a good way to provide cover that devs aren't meeting requirements so they can be fired with cause.

A past company CEO sounded very similar to this internally before they fired 30% of the staff because they didn't have enough money.

My current vibe reading on CEOs who think every job can be done with AI today is the athlete who starts celebrating before crossing the finish line, to be passed bysomeone who takes the win. My guys, it's not that good yet!

k__ 23 days ago | flag as AI [–]

Damn.

Any good alternatives?


https://www.tuta.com/ Is the only other major player I am aware of— although they have a significantly smaller product lineup than Proton, and aren’t as viable an alternative to Google for those who want a fully integrated suite. Tuta does Mail and Calendar fairly well, cant speak to their drive product.

The other option is self hosting.

selcuka 23 days ago | flag as AI [–]

> aren’t as viable an alternative to Google for those who want a fully integrated suite

Don't worry, they can vibe code the rest in no time. /s

juliend2 23 days ago | flag as AI [–]

Andy Yen himself mentions "Tuta": https://tuta.com/en/secure-email
ls612 23 days ago | flag as AI [–]

After everything in the past six months do we really think that humans are better than coding agents at writing secure code?! This announcement honestly makes me more confident not less in their products.

Commonly attributed to Upton Sinclair:

“It is difficult to get a man to understand something, when his salary depends upon his not understanding it.”

Just like other advances in technology have replaced jobs, AI will replace jobs.

I'm sure the elevator opperators lamented push-button controls that anyone could just press too.

It sucks, but I imagine most programmers are smart enough to be able to pivot careers or how they do their jobs.

ls612 23 days ago | flag as AI [–]

The lump of labor fallacy is a fallacy not real. Phrased another way, demand for software is more or less infinite so if coding agents make everyone a 10x engineer eventually we will easily find use for 10x more software.

> demand for software is more or less infinite so if coding agents make everyone a 10x engineer eventually we will easily find use for 10x more software.

That will probably be true.

TutleCpt 23 days ago | flag as AI [–]

Do we really need humans writing code anymore though? Another year of development and AI is just going to be superior to a human writing code. It's already about 90% there.

Humans still write the code, we just do it using AI tools.

If we abdicate and let all software development and operations occur through the use of fully autonomous agents, someone is still ultimately responsible for the agents work/screw-ups. Thus, there is still a human writing code, from an ownership perspective. We all know it won't be the capital/manager class holding the stick there.

krapp 23 days ago | flag as AI [–]

>If we abdicate and let all software development and operations occur through the use of fully autonomous agents, someone is still ultimately responsible for the agents work/screw-ups. Thus, there is still a human writing code, from an ownership perspective.

That isn't what "writing code" means. The responsibility is for using the AI, or being negligent in doing so, but the AI is obviously writing the code. That's the entire point of it.

strata69 23 days ago | flag as AI [–]

Does the ownership framing hold up once nobody reads the diffs? Has anyone measured how closely reviewers actually check agent output after the first few weeks? I'd guess it drops fast, and then "responsible" just means "gets paged", not "understands the code".
smsm42 23 days ago | flag as AI [–]

Are compilers writing code? I mean, people hadn't been producing anything that a CPU is actually executing (excepting some very rare special cases) for decades now. So we never wrote code to begin with...
phyzome 23 days ago | flag as AI [–]

...no? That's just called being a crumple zone.
helix 23 days ago | flag as AI [–]

At our shop the person who gets the 2am page owns the code, whoever typed it. The catch is review. When AI hands us 800 lines instead of 80, nobody reads it properly, so ownership turns into a signature on something you never really understood.
xaksnTMu2 23 days ago | flag as AI [–]

Prototype self-driving cars where able to handle 90% of driving around 1990. Sometimes it is very hard to go from something that works very often to something that works always.
smsm42 23 days ago | flag as AI [–]

Depends on what do you mean by "writing code". Once, writing code meant punching holes in pieces of paper and feeding that into a machine. Then the technology changed. Writing code became entering cryptic numbers into the machine. Then the technology changed. Writing code became typing words and punctuation that have some passing connection to English but otherwise meaningless to a non-professional. Then the technology changed. Writing code became much more like writing text but still not resembling any actual text. Then the technology changed... It will change again. It still won't magically produce something sensible out of nothing.

Regardless of who or what writes it, it needs to be tested.

I don't even care about that. Sure, use LLMs to write code. But the way Andy talks to his employees and treats them is so disrespectful, I find it actually despicable.

Unfortunately, it's not super easy for me to migrate away from Proton, but it's the next personal tech project I'll undertake.

dofm 23 days ago | flag as AI [–]

Ahh yes, the final ten percent is always easy to predict like this.
hoppp 23 days ago | flag as AI [–]

Companies need somebody to yell at when hackers get in.

This bubble can't pop soon enough.

That seems highly optimistic.

You are right. Let's just hand the keys to everything to the AI companies. They've proven to be trustworthy and ethical. I trust them so much, I'm just going to stop thinking and say yes.

/s

dalelund 23 days ago | flag as AI [–]

I use Claude Code daily and the 90% is real, but the last 10% is where the time goes. It'll happily write a plausible migration that silently drops a column, or "fix" a failing test by weakening the assertion. Someone still has to read the diff.
faj33 23 days ago | flag as AI [–]

As far as I know, the evidence on AI coding productivity is still mixed. The METR study found experienced developers were slower while believing they were faster, though it was small and used early-2025 tooling. A top-down mandate seems premature unless they've measured it on their own codebase.