I would add AI if I can get people to pay for the regular version as it is (where people just manually tag their bookmarks). Although with the filter, you can filter by a keyword and tag a bunch of them at once.
X is Twitter. You bookmark practically anything (post, comment, quote) and can find it later in one place.
I did put it together on a whim because I don't trust that something won't happen to X.com some years in the future with Elon and you legitimately cannot export your bookmarks from X. If that happens, your bookmarks will be lost.
Yes, it is a paid product because I put serious time into it and I designed it to be something I would use myself (even if I make no money off it).
Features are really simple. The annual license is $4.69 because I expect people to export their bookmarks only once every few months.
1. Chrome extension to scrape your bookmark page. You can find plenty of extensions that will do this, but they give you a JSON/CSV file that's not very useful.
2. The viewer is paid. You import the JSON from Step 1 and it loads all your bookmarks.
3. You can tag your bookmarks. You can download the media files embedded into each bookmark. You can delete your bookmarks.
4. You can export your bookmarks as another JSON that you can save or give to a friend (if you trust them with your private data).
5. I may add AI tagging for Claude to suggest tags for your bookmarks.
The Chrome extension exports everything on your X bookmarks page to a local file. No login, no data leaving your machine.
If your account gets suspended tomorrow, your bookmarks are gone. This prevents that.
I also built a viewer if you want to search, tag, and browse your exports like an actual library — offline, no account needed. $4.69/year for the viewer, but the exporter is fully standalone if you just want the backup.
It uses only React. If you want to support me, it costs almost nothing. Advice also welcomed.
How do you assess the consumer value proposition? I think everyone would want to know if the images and videos they find on social media are real, but who would find it essential to know and would be willing to pay a service like yours to get that answer?
Amidst a lot of analyses and results I can vaguely understand, this conclusion stands out:
We assess that Claude Mythos Preview does not cross the automated AI-R&D capability threshold. We hold this with less confidence than for any prior model. The most significant factor in this determination is that we have been using it extensively in the course of our day-to-day work and exploring where it can automate such work, and it does not seem close to being able to substitute for Research Scientists and Research Engineers, especially relatively senior ones. Although we believe this is an informed determination, it is inherently difficult to make its basis legible, given the model’s very strong performance at tasks that are well-defined and verifiable enough to serve as formal evaluations.
The ECI slope-ratio measurement we introduce in section 2.3.6 shows an upward bend in the capability trajectory at this model, though the degree of the upward bend varies significantly across dataset and methodological changes we made to stress-test it. The identifiable driver traces to specific human research advances made without meaningful assistance from the models then available. That said, we will be continuing to monitor this trend to see whether acceleration continues, especially if this is plausibly traceable to AI’s own contributions.
The bottom line: This new Claude model is not yet capable enough to autonomously do AI research — but it's closer than any previous model, and Anthropic is nervous about it.
What's the "automated AI-R&D capability threshold"?
Anthropic has defined a danger line: if an AI can independently do the work of AI researchers, that's a big deal — because then AI could start improving itself without humans in the loop. This assessment is asking: has this model crossed that line?
Why are they less confident than usual?
With past models, the answer was a comfortable "no." This time, they're saying "no, but..." — it's a much closer call. They're hedging.
The AI researchers designed tests to evaluate whether the model can do their real day-to-day work. They found out Mythos scored well on structured tests, but they know themselves that structured tests do not capture the non-linear, intangible aspects of AI research. So, interesting results, but AI can't replace them yet and AGI still far away.
Man could have kept low and build a nest egg until he gets busted by a functioning federal investigatory agency, but no, he has to stroke his ego and get a NYT profile and now his fraud is in the public eye.
Individuals with irregular bedtimes had a 2.01-fold higher risk of MACEs compared to those with regular bedtimes (HR = 2.01, 95% CI: 1.00–4.01, p = 0.049), and those with irregular sleep midpoints had a 2.00-fold higher risk compared to those with regular midpoints (HR = 2.00, 95% CI: 1.01–3.98, p = 0.048).
So it doubles one's chance of a heart attack or stroke. Gnarly.
It would be useful if you explain how you calculate it. I mean, if you just apply a decaying exponential function, anyone can do that on their calculator.
Most of us should be honest and admit we're jealous of this guy.
He basically chose a sector where customers are desperate (weight-loss drugs), slapped a website and an interface for connecting with a drug prescription provider together, did effective marketing, and now his business generates millions a month in profit.
Like, there are a half dozen companies like his running around that essentially offer the same product and prices because they are all customer interfaces stop the same provider.
Just the other day I was downvoted and called out for suggesting that perverse incentives are hard to resist, yet here we are with the Times (apparently) showcasing another such instance.
In this case GLP1's clinical effects are widely understood though, so it is immaterial if an "artist's depiction" (artificial agent's depiction) is of a real person or purely hallucinated.
This is just like when Paypal got started and was basically operating their own bank. Good luck doing that without getting in trouble. This is selling pharmaceutical drugs over the internet. You're playing chicken with going to jail they just happened to get lucky.
reply