I spent half my day following your launch, and it looked like things were going downhill pretty quickly. How have things been since? You turned on your paid subscription signup?
Years ago I tried to install and sign up for Turo on iOS to rent out a car I owned. It was a luxury car with a rebuilt title.
After I put in the VIN of the car, I received an error, and inexplicably I was banned from the app. No notification as to why, no "we don't accept rebuilt title vehicles," nothing. Naturally I scoffed, deleted the app and forgot about it.
Last year a friend rented a few cars on Turo for a trip and added me as a driver to one of them. I had switched phone numbers but kept the same phone. I downloaded Turo again and signed up with a new phone number and new email.
Before Turo even asked for my driver's license information, I was blocked again. It must be due to fingerprinting, which persisted over years.
I'm unsure how much apps can learn about your user profile, other apps you have installed, and other uniquely identifiable data. I've assumed it was limited, but perhaps I've been naive.
I guess these new rules are generally good? But I can imagine for every nefarious usage of these APIs, there can be a plausible cover reason...
I heard recently (on Lex Fridman's podcast with Marc Andreessen) that Oppenheimer was called a "cry-baby scientist" by President Truman, and that his groveling campaign was considered by many to be self-aggrandizing, albeit in a self-deprecating way.
Apparently, his involvement was less significant than many others in the Manhattan Project, but his name became one of the most well-known due to his public outcries, and of course his iconic quote.
I haven't yet seen the movie and I'm looking forward to it; I wonder how much of this is portrayed.
Funny enough, I'm planning a double feature with Oppenheimer and the Barbie movie. Maintaining a healthy relationship involves tradeoffs.
Cool, I've had "add a personal calendar appointment scheduler feature" on my backlog for months, and for some reason I just didn't want to do Calendly. Maybe it's because I'm pretty averse to overgrown SaaS?
I like their messaging and I'm more inclined to use a tool ran by a team/company like these guys.
Consider me a customer... whenever I clean up the backlog
Surprised to hear no mention of Obsidian (https://obsidian.md/). I switched from Evernote to Notion years ago and finally switched beginning of this year from Notion -> Obsidian.
Top reasons:
- local first files stored in markdown format (somewhat future-proof); this also allows a Git repo sync
- multiple ways of syncing Vaults (their term for a workspace); can store Vaults on iCloud, Dropbox, or with Obsidian's Sync service which is E2E encrypted
- rich plugin ecosystem
- can style pages with CSS and run various code, like javascript, from inside the editor
- Graph view and Canvas are extraordinarily useful visualizations of your data
- supportive and prolific community, from their forums, to YouTube creators and bloggers; simply search "Obsidian intro" or "Obsidian tips" or "Moving to Obsidian from ___" for inspiration
It seems like there's a lot of recent interest and effort in open-source or self-hosted Notion-like/markdown-with-widgets applications and platforms. AppFlowy (https://github.com/AppFlowy-IO/AppFlowy) comes to mind; I attended one of their monthly "town hall" meetings a few months back, and looks like they're rapidly increasing in popularity. I think there was another similar project like this on HN front page last week, IIRC.
This makes me happy, because I switched to Obsidian primarily for local-first file storage in a platform-agnostic format. I've learned to love many things about Obsidian and am writing a few plugins myself, but there are still several Notion-esq functionalities I wish I had, and I find myself handing off between Obsidian and other webapps for certain effort, like team project management.
I used to get far more excited to explore new projects like BlockSuite, and I really appreciate their documentation, but I find it hard to justify allocating time to reviewing and trying out new tools when I still have much more improve on with my Obsidian usage; this is especially true of newer projects where I'm unsure of their shelf life.
To assuage my internal conflict I remind myself that I think plaintext is fundamentally the right choice for much knowledge collection, and I'm proud to say that if the internet shut down, I'd retain a significant growing fraction of my personal data.
So, to mitigate, avoid changing Github organization names if possible, and if not, preserve / re-register the former organization name to prevent impersonation?
My first 5 minutes have been positive. I've worked for years on a similar platform but never expected to have the time to ship, so I appreciate that you've implemented a few patterns/features that I was hoping for, namely tag-based subscriptions.
Bug report: on latest Firefox with uBlock Origin, I have a strange UI bug on the root/home page: without any activity from cursor or otherwise, there is a dropdown that appears to show emails saved in my browser for form auto-fill. This is visible in the top left which seems to be in the same place where your registration sidebar is.
I have quite a number of thoughts and questions which I'll move to the platform. Great job.
I pretty strongly disagree with this take. I’ve used Reddit for over a decade almost entirely as a “lurker” meaning without an account. The majority of what I watched and read has depended on that reaching the “front page of the internet.”
While I’ve followed these recent changes since the Pushshift shutoff, it’s really only been in the past week that I’ve seen this gain “mainstream” attention. The more of these subreddit-going-dark announcements reach the page, the more people realize this is a widespread issue - even if they don’t click, read, or even think about what’s going on or why, they’re aware of it.
I’ve seen several dozen of these posts reach the front page, plus a litany of media outlet publications. That legitimately matters as an input to financial projections in the context of their upcoming IPO, and it matters even more as a signal to prospective investors, retail or institutional.
Our modern reality is often served to us via algorithms that are trained to optimize for things; the most common thing for content platforms to optimize for is engagement, and the most common signal for these platforms is the upvote. In over a decade I’ve never seen any topic reach front page across such a variety of posts and subreddits, and that is a direct result of individuals upvoting these posts.
I think this demonstrates, already, the collective ability to mobilize action across communities, even if the action is reduced to the simplest Boolean upvote. And that’s an indication that widespread collective action can be coordinated.
It's funny that I can think of a few legitimate use cases for this type of service:
- Deploying a new brand identity. I've worked with larger marketing agencies that may create many dozens of brands in a given month, and the process tends to be fairly manual and subject to various operational issues. There may be 20+ usernames/handles to "claim" as accounts across a broad spectrum of platforms, plus updating bios, profile pictures, 2FA/MFA, and, of course, the basic sign up and email confirmation process. I once scoped out how much of this could be automated, which was unsurprisingly cost infeasible relative to the current human method, and one of the components involved verification links, especially in scenarios with multi-tenant brands (such as a client wanting their new socials to be established under their own email, but wanting the agency to create those socials).
- User software provisioning. For all services that don't support some type of OAuth provisioning (through Okta, Atlassian, Google, etc) it would be helpful to create a new user, use scripts to post sign-up requests to non-OAuth services, use a function to verify the user's email, archive the email and present the user with their credentials.
- Tracking marketing/sales/promotion/update/announcement emails. Imagine if I could pay per user record for a Regex/SQL query; what if I want to estimate the amount of traction or activity of a competitor's email list? Rather than facilitating any interaction with the source users' email, it could simply give me a count of records, and perhaps a boolean for read/unread. Already that sounds like valuable data.
I think it's very clear that the recent LLM boom is directly responsible for Twitter, Reddit, and others quickly moving to restricted APIs with exorbitant pricing structures. I don't think these orgs really care much about third-party clients other than a nuisance consuming some fraction of their userbase.
Enterprise deals between these user generated content platforms and LLM platforms may well involve many billions of API requests, and the pricing is likely an order of magnitude less expensive per call due to the volume. The result is a cost-per-call that is cost-prohibitive at smaller scales, and undoubtedly the UGC platform operators are aware that they're pricing out third-party applications like Apollo and Pushshift. These operators need high baseline pricing so they can discount in negotiation with LLM clients.
Or, perhaps, it's the opposite: for instance, Reddit could be developing its own first-party language model, and any other model with access to semi-realtime data is a potentially existential competitor. The best strategic route is to make it economically infeasible for some hypothetical competitor to arise, while still generating revenue from clients willing to pay these much higher rates.
Ultimately, this seems to be playing out as the endgame of the open internet v. corporate consolidation, and while it's unclear who's winning, I think it's pretty obvious that most of us are losing.
I worked with an electrical engineer who disclosed to me a mid 8 figures investment made by a PE firm to develop Ethereum FPGA and ASIC hardware, and while he told me they were underwater on that deal, they did achieve (IIRC) 30%+ performance (watt to flop) relative to (at the time) top-of-the-line GPUs.
I wonder what they’re doing with that hardware now.
Depending on the Terms of Service flavor your platform of choice serves you, it’s generally that you own the content you post, but grant an infinite, irrevocable license to use that content for just about any reason, including as an entry in a dataset sold to a separate entity for use training a large language model.
I think this is a really important question worth answering by future UGC-based platform incumbents.
You’re right, but I think it’s also pretty clear that
A) there is demand for functionality that depends on semi-real-time data, e.g. a prompt like “explain {recent_trending_topic} to me and describe its evolution” where the return could be useful in various contexts;
B) the degradation of search experience and the explosion of chat interfaces seem to indicate “the future of search is chat” and the number of Google searches prefixed or suffixed with “Reddit” make it obvious that LLM-powered chat models with search functionality will want to query Reddit extensively, and in the example prompt above, the tree of queries generated to fulfill a single prompt could be sizeable;
C) improvements to fine-tuning pipelines make it more and more feasible to use real-time data in the context of LLMs, such as a “trending summary” function that could cache many potentially related queries from Reddit, Twitter, etc and use them to fine-tune a model which would serve a response to my example prompt
I think it's very clear that the recent LLM boom is directly responsible for Twitter, Reddit, and others quickly moving to restricted APIs with exorbitant pricing structures. I don't think these orgs really care much about third-party clients other than a nuisance consuming some fraction of their userbase.
Enterprise deals between these user generated content platforms and LLM platforms may well involve many billions of API requests, and the pricing is likely an order of magnitude less expensive per call due to the volume. The result is a cost-per-call that is cost-prohibitive at smaller scales, and undoubtedly the UGC platform operators are aware that they're pricing out third-party applications like Apollo and Pushshift. These operators need high baseline pricing so they can discount in negotiation with LLM clients.
Or, perhaps, it's the opposite: for instance, Reddit could be developing its own first-party language model, and any other model with access to semi-realtime data is a potentially existential competitor. The best strategic route is to make it economically infeasible for some hypothetical competitor to arise, while still generating revenue from clients willing to pay these much higher rates.
Ultimately, this seems to be playing out as the endgame of the open internet v. corporate consolidation, and while it's unclear who's winning, I think it's pretty obvious that most of us are losing.
Sad to see no mention of the Motorola Q (1) which I got in 2006. My first mobile phone was also my first ‘smartphone’ with a full keyboard. I didn’t have any cellular service (I was a child) so I downloaded a ~2gb package with 350,000 text-only Wikipedia articles onto a MiniSD card (before microSD!), and that kept me endlessly occupied.
I still have that device, quite possibly my oldest possession. Maybe one day I’ll find a way to get it working.