epoch_100

> DeepSeek does not "do for $6M what cost US AI companies billions". I can only speak for Anthropic, but Claude 3.5 Sonnet is a mid-sized model that cost a few $10M's to train (I won't give an exact number). Also, 3.5 Sonnet was not trained in any way that involved a larger or more expensive model (contrary to some rumors).

Wow!

Anthropic’s privacy policy is extremely strict — for example, conversations are retained for only 30 days and there’s no training on user data by default. https://privacy.anthropic.com/en/articles/10023548-how-long-...

Very cool! It reminds me of https://chord.pub/.

I use Anytype and like it. I do hope they’ll invest more time into the editing experience — there are a few sharp edges that will probably alienate people coming from more mainstream tools like Google Docs and Notion:

- Typography is a bit weird, especially with headings; there is very little line height on headings, leading to cramped documents.

- Bullet points look weird when you look closely at them. (On Mac, they’re not perfectly circular; perhaps an aliasing issue.)

- Can only use cmd+b and cmd+i for bold and italic when selecting text; can't use them to switch on/off boldness on the cursor.

- Can't specify type when creating a new object with the @ symbol inline (Capacities, a related tool, does).

- Bullet points flicker when you first create them.

- Linking inserts a space even when you don't want it, as does formatting using markdown characters (e.g., providing the closing ** in italics.)

- Relatedly: Entering closing markdown characters (e.g., the second **) does not end the formatting effect (the cursor should not have the formatting applied after they're entered, but it does).

Anytype is overall a fairly polished app. I hope they’ll iron out these editing issues — when they do, it’ll really be exceptional.

Author here. I agree. I designed a17t awhile ago (I think nearly four years ago now?) and it’s due for a refresh. I think it has the potential to be excellent but it’s not there yet.

Really what happened is we checked whether we could set `isAdmin` to `true` on our existing accounts, and... we were able to. Adi's more technical writeup has details: https://saligrama.io/blog/post/firebase-insecure-by-default/

Adi's writeup is great, and goes much more into the technical detail than my transcript. I really recommend everyone checks out his post.

I am not a lawyer (of course). But I find some solace/comfort in the new Justice Department guidance to not charge good faith security researchers under CFAA. https://www.theverge.com/2022/5/19/23130910/justice-departme...

Yes, threatening to report is what was really happening here. But in their effort to scare us, they elided much of that process. From our perspective it was "watch out, you might face felony charges if you don't agree to silence".

Heh. That's not very nice. If it feels stilted, that's probably because I wrote this primarily to be spoken. But ChatGPT was not involved.

> but you cannot distinguish between limited access and full access

I believe you can check whether you received limited photos access: https://developer.apple.com/documentation/photokit/phauthori...

Shynet

Interesting — what kind of advantages would your hosting service provide over, say, packaging the Pynecone app into a Docker container and then deploying to Fly.io/Render/GCP/Azure/etc?

We hacked that together for https://paxo.ai — can be done!

If the VCs haven’t made a capital call, then that unallocated money is likely still sitting with the LPs collecting that sweet risk free return.

Ha. I’m also doing something similar with a friend at https://www.paxo.ai. Funny that we all seemed to have an similar idea, all at once.

I like them. But the outages have been tough.

There’s https://atlos.org for visual investigations

Shynet: https://github.com/milesmcc/shynet

The goal is to provide modern, privacy-friendly, and detailed web analytics that works without cookies or JS. And it's completely open source.

Full disclosure: I am the primary maintainer.

Author of Shynet here! Hope you like it :)

I have a strong technical background and a passion for digital safety and privacy. I'm especially interested in trust & safety, privacy engineering, human-centered design, tech policy, and open source software.

I'm looking for an internship or fellowship adjacent to trust & safety / privacy engineering.

Some of the projects I’m most proud of are Shynet [0], PrivacySpy [1], PolitiTweet [2], and a17t [3]. I co-instruct CS 106S [4] at Stanford, and I worked on cyber policy for a 2020 presidential campaign. I also work at the Stanford Internet Observatory on both research and technical infrastructure.

Portfolio: https://miles.land

Location: NYC or SF Bay Area

Remote: yes but ideally no

Willing to relocate: no

Technologies: Python, Rust, JS, Java, Kubernetes, C, OSINT, web dev, and more.

Résumé/CV: https://docs.google.com/document/d/1wnIBWjEPmgdYXQ_EYZRU--KE...

Email: [email protected]

[0] https://github.com/milesmcc/shynet

[1] https://privacyspy.org

[2] https://polititweet.org

[3] https://a17t.miles.land

[4] http://cs106s.stanford.edu

I have a strong technical background and a passion for digital safety and privacy. Especially interested in trust & safety, privacy engineering, human-centered design, tech policy, and open source software. Looking for an internship or fellowship adjacent to trust & safety or privacy engineering.

Some of the projects I’m most proud of are Shynet [0], PrivacySpy [1], PolitiTweet [2], and a17t [3]. I co-instruct CS 106S [4] at Stanford, and I worked on cyber policy for a 2020 presidential campaign. I also work at the Stanford Internet Observatory on both research and technical infrastructure.

Portfolio: https://miles.land

Location: NYC or SF Bay Area

Remote: yes but ideally no

Willing to relocate: no

Technologies: Python, Rust, JS, Java, Kubernetes, C, OSINT, web dev, and more.

Résumé/CV: https://docs.google.com/document/d/1wnIBWjEPmgdYXQ_EYZRU--KE...

Email: [email protected]

[0] https://github.com/milesmcc/shynet

[1] https://privacyspy.org

[2] https://polititweet.org

[3] https://a17t.miles.land

[4] http://cs106s.stanford.edu

Author here. It's to encourage people to apply, and to bring attention to what I think is a really awesome place. I don't want it to come across as pompous, though I get that to some readers it'd come off that way. :/

I also feel a bit obligated to include this, because I got a _lot_ of feedback on this piece from other Recursers — if it wasn't for them, this post probably wouldn't exist!

It's absolutely, unequivocally better that these tools be open source (and therefore fixable!) than closed source. My (admittedly small) gripe is just that a large, well-resourced corporation is shifting the maintenance burdens of its product to customers. This situation is certainly better than having to pull out the debugger or beg with customer support.