This is a good point. AI vibe-coded stuff that hasn’t been pentested being pushed in a product promising security and privacy is genuinely terrible. Perhaps that one thing is so terrible it should rightfully overshadow everything else because of the risks involved.
Are you sure? The privacy policy says “This site is hosted on GitHub Pages by GitHub, Inc. (USA). When you visit, GitHub logs your IP address for security purposes (Art. 6 (1) (f) GDPR: our legitimate interest in a secure, working website). The site cannot be delivered without it. GitHub keeps these logs as long as it needs them for security and publishes no fixed period. GitHub takes part in the EU-US Data Privacy Framework; see the GitHub Privacy Statement.” I think the server itself may actually be GitHub somehow?
I agree with this. I like AI and really don’t like people labeling anything AI-created as slop. But in order for something to not be AI slop, the developer needs to review the output and provide feedback to the model. If this is “draft 1” of the content and the dev only reviewed parts of it, I understand dev’s excitement and wanting others to see it, but it seems like dev should be beta testing draft 1 with the model and doing some revisions before asking others to check it out.
Because some of the phrasing is unclear enough in its explanations that a typical person might not be able to follow it. And I feel like if dev had read more of it, dev probably would have noticed that…
How can it not be on a server? Does that mean github hosts it and you just own the domain? (I am reading your privacy policy.)
So it involves both Github and Cloudflare, which means really not only is an IP being logged, but also other things like packet latency, mtu, packet header characteristics, and an overall network fingerprint.
This is also the same Microsoft that helped use Microsoft Azure servers for various war activity, so this does mean data about anyone marginalized who uses this will be captured and used however… Most people don’t care about data privacy, but still it isn’t ideal.
It doesn’t look like you’re a moderator, according to the sidebar. How would you be able to remove it if you wanted to do that? I am not being snarky. Genuinely I’m confused.
I agree with you that the biggest weakness of the project is there isn’t enough clarification of conceptual topics and there’s no place to get more info on things. There’s a suggest an improvement flag also, but it’s in a strange place (not near the topic/question) so it’s a bit unintuitive for suggesting the conceptual explanations are weak.
But I also think that if the person who made this project wanted to try to correct that, they probably could. I hope the developer answers some of your questions. I think the “What are the chances you’ll abandon this project?” is a valid question. Because this project needs work to be more user-friendly and if people give feedback and it doesn’t end up being implemented, then the people giving feedback were wasting their time.
Impovrished people are exactly the sort of people who may need an open source educational tool to get ahead. Lots of people are impovrished.
I like it! But I probably wouldn’t use it right now because I’m studying other things.
I did test it briefly.
My experience was that for subjects I understood already, I could do the tests and follow along easily.
Then I tried Electrochemistry & Batteries.
So a few things. First of all, I can’t easily go back and see the work I previously did, so I am doing this from memory. I could go back with a reset. I’m not sure if that’s a good user experience that I can’t go back to review previous ideas and concepts.
Oh wait, now I am looking at the course and it started over. I didn’t ask it to reset my progress. Hmm. That may be a bug?
One thing I didn’t like is that if I didn’t understand a concept, I couldn’t easily get more information on it or additional help.
For example, in this prelude to a question:
Oxidation numbers
For molecules, chemists track electrons with oxidation numbers: pretend every bond is ionic, giving the shared electrons to the more electronegative atom. Rules of thumb:
free elements: 0; simple ions: their charge, oxygen is usually −2, hydrogen +1 (with non-metals), the numbers add up to the overall charge.
An increase in oxidation number is oxidation; a decrease is reduction.
I was confused by this. I haven’t studied chemical bonds in a long time. I don’t know why I would pretend the bonds are ionic, and I don’t really have memorized which atoms are more electronegative. So I read this and sort of get it, but I’ve also studied this a long time ago…
Now for the question:
Manganese in permanganate What is the oxidation number of manganese in the permanganate ion, MnO4−MnO4−?
Okay… so we just went over oxidation numbers. I have to pretend there are ionic bonds (I don’t know what that is exactly from memory, I know it’s a type of chemical bonding). And electrons move to the more electronegative atom. Electronegative atoms wasn’t exactly explained clearly here. I don’t really know how to determine if Mn is electronegative. I see O4 has a negative next to it. Is this negatively charged Oxygen? Does that imply negative electrochemical properties?
I didn’t know, guessed, got it wrong, and then saw the answer:
Explanation
x+4(−2)=−1x+4(−2)=−1, so x=+7x=+7, manganese’s highest oxidation state. That’s why permanganate is such a powerful oxidising agent: manganese is hungry for electrons.
This doesn’t make sense to me. I know the previous screen said -2 was often a number associated metal (“oxygen is usually −2”; it also didn’t clearly define what that is. Is that the charge? Is that the oxidation number?)
After that, the next thing is:
Step 5 of 6 Try it Balance a half-equation In acid, purple permanganate is reduced to colourless Mn2+Mn2+. Balance the half-equation: oxygen first (with water), then hydrogen (with H+H+), then charge (with electrons).
I was able to figure out the right answer by trial and error, but I don’t actually conceptually understand this.
So my main feedback is that this may help people who have studied things before or want to look up things on their own, but for people who don’t understand terms or need remedial help on a concept, there’s not really a way to reinforce lessons or clarify topics.
I don’t know if every lesson is like this, but the one thing that could probably help this the most would be something like a button for “I still don’t understand” that has longer explanations, more examples, and more similar problems. I’m somewhat flexible when learning and don’t need to understand everything right away… but I need to be able to get conceptual clarification on something before trying to build on weak conceptual understandings when I’m studying a topic.
I am sure I could step out of the ecosystem and just ask an LLM why I am not understanding certain parts, but that definitely takes me out of the flow.
I am also not trying to re-learn chemistry or batteries right now. (But I’m pretty sure I never learned how lithium ion batteries worked and so if I could have followed this lesson and then learned more about new battery technology, I would have found that interesting. But I had to stop because of conceptual confusion, and I don’t want to take the time to have an LLM fill in the gaps. But for me to use this, I would probably need to be able to log in, have my progress saved, and harder modules might need more links to various terms. I did skip ahead and I know that I could have covered chemical reactions first. I did try part of that module and it felt too easy.
For me, if I wanted to re-learn chemistry right now so I could learn more about science and improve my science abilities, I would likely just have an LLM teach me directly because I think it could just be faster.
For people who might be interested in using this, for example, a teenager who wants to take some community college courses in lab work because they want to become a genetics researcher and they want to pre-learn some of the material so the course will be easier later, I just don’t know if the explanations are long enough or clarify things enough. Those explanations felt really sparse. I like the structure, I like the diagrams and pictures and those are nice. But just for anything slightly harder, there’s no real clarification when there are conceptual misunderstandings.
You could probably create links for terms (like ionic) that help people review what those things are or their definitions. And you could probably create longer deep-dives for people who get a problem wrong or don’t feel comfortable moving on and then it would work better.
Best of luck with this project!
I sort of get where he is coming from. AI has made shockingly fast rapid advances in being able to code. AI code should definitely be pentested, even if it’s just with more AI.
There are many developers and tech privacy enthusiasts who hate AI coding for whatever reason, and I get his perspective. He just isn’t being very strategic because he’s acting like a normal CEO and not caring that much about emotional impact because he’s in a leadership position, forgetting that a privacy-focused user base will not tolerate typical CEO behavior.
I already have real fears regarding Proton. What sort of privacy company doesn’t accept Monero??? What sort of privacy company uses US-based Zendesk which could be required to cooperate with a police state with a secret dragnet warrant??? Proton has a very active fraud/abuse detection system. Even if it is necessary, how could that be technically possible if messages are encrypted???
But we should also remember that if you are a poor person in a repressed regime, you can get free Proton services to help get you more access to information. They are very nice about that. Who made that choice so that a poor person in a poor country can get these services for free if they are poor? Andy Yen made that decision.
People make mistakes and if Andy Yen realizes he needs to be more emotionally intelligent and aware and tries to change, perhaps people should be willing to forgive him. Proton does a lot of good overall despite recent concerning behavior.
I will also add that it’s concerning that their AI Lumo is integrated into so much potentially and it’s not clear where the processing occurs, if the processing itself is encrypted during the processing, and what models are used. For example, if Lumo is using US-based company data centers in the US and the model has to be given unencrypted data to process, then in theory if there’s a gag-order demand on the data center, all of that information can be extracted. I see very little information about that, which is concerning.
Once they started rolling out identity verification and KYC social media, with age checks and Persona and the like, it made the decision easy and that was the end of it for me.
The link between big tech Orwellianism and ad tech Orwellianism and police state Orwellianism is too great. I won’t be a frog in a pot. Done with all of it.
https://discuss.privacyguides.net/t/cbc-tutanota-tuta-is-a-honeypot/14978/9
Why would a government employee, being charged with a crime, randomly make up lies about tutanota, now Tuta?
The likely answer is he didn’t make up lies and wanted to expose an operation as revenge for them charging him with a crime, which he thought he would likely be convicted of.
Cameron Ortis did actually get 14 years in prison. So if this were merely a disinformation campaign, why did he end up in prison?
For this accusation to be false, you need it to be true that Ortis randomly decided to lie about Tutanota (now Tuta) during a trial or had some incentive to lie during trial to spread misinformation about Tuta. However, there was no plea deal in this case. He was granted day parole in September of 2026, but I don’t know that much about Canadian prisons or law and if this would be considered unusual.
The fact that Tutanota rebranded to Tuta nearly immediately after this (to make it harder for regular people searching for this to discover it) is also suspicious.