curious_curios

Put your site behind cloudflare with geo and bot blocking should eliminate the vast majority of them.

Do note there’s nothing stopping you from doing both.

As I recall it was primarily based on some rough hardcoded heuristics on your searches. It could be reliably reproduced with certain exact searches, and seemingly only targeted a few programming languages.

Definitely something to this. I think that’s why so many social media platforms are entirely image/video based because that’s easier on the phone.

Some approaches we’ve used:

- Group the comments by theme, then pass groups to the LLM to summarize/deduplicate then pass the outputs of that as context.

- RAG where only relevant parts are included in the context.

- Use an LLM with a larger context window (like Gemini pro)

You’re welcome!

Hit up the freelancing firms staffing your company, they probably work for others in different industries too. Note they’ll take their cut, but you’ll probably still get paid more than your current hourly.

There’s a whole world of unintuitive reasons for this kind of thing. Non-exhaustive list:

- Freelancers are OpEx, FTEs are CapEx. They are usually separate budgets so departments split.

- They want to maintain good relationships with the freelance companies in case they need to spin up staffing quickly with top talent in the future.

- They don’t have to pay some benefits and taxes on freelancers.

- Freelancers and consultants are often used to get tabs on ‘industry best practices’, i.e. what your competition is doing.

- Despite what you might think, they might actually be better than the FTEs, or at least perceived to be so.

- It’s easier to fire them when times get tough without as many rules and regulations.

NAL but we’re in uncharted territory here with the administration ignoring court orders.

When AI outputs have extensive primary source citation and are peer reviewed then sure. Even now with deep research and web browsing capabilities it doesn’t vet its sources though. It has a long way to go.

Historically the moderation tooling on Reddit has been overlooked and terrible. There was a lot of tooling that got built up on the Reddit API to help with this, but a lot of that got torched when they locked down the API. When you add in how easy it is for scammers and spammers to use gen AI to overwhelm moderators it’s really not a great experience moderating a community.

At this point moderators are aiming for literally any filter they have to make their task possible, you just got caught up in this arms race.

Likely yes, via TURN servers.

It’s largely dependent on the client’s bank/card issuer, but it’s definitely well under 50%

I assume there’s payment processor reputation that they get dinged on that they want to price in.

Don’t fight them, just block the user from ever using the service again and keep going. Statistically you’ll lose even with airtight evidence.

Documents don’t take up much space. Videos absolutely do though, so you’ll get a pretty good return on investment in going through those.

These questions are asked (at least in the US context) due to required reporting to the EEOC: https://www.eeoc.gov/data/eeo-data-collections

The general consensus is that political posts are not viable on this site. I interpret that anecdotally from the fact that any post involving anything politically adjacent, including posts about technology related to a person in the political sphere, are quickly flagged. The discussions that happen prior to being flagged are largely charged, talking past each other, regurgitating points and fallacious.

From what I’ve seen the recommendation is to go to sites like https://www.themotte.org/ to have these discussions.

I think there’s potential for a race condition between when you close the Listener and the http server spins up on the no longer reserved port that we’ve seen with lots of tests being run in parallel. You can return the listener, use t.Cleanup to close it and serve your web server on it.

It’s time to start hiring people. Seriously, you will not regret it.

3rd party attestations tied to secure hardware enclaves.

I get the sentiment, but I can say without any hyperbole that would be a death sentence on the majority of people with chronic conditions.

The VP is looking out in their best interests, the IC should do the same and move to the new team on the original schedule. Any promise not in writing will disappear most of the time, even then it can be walked back.

As best as I can tell it sounds like the emails are coming from _within_ the NOAA network, not your usual spam.

I think this is a result of the Chatham House Rule, which is very much in vogue in the startup world these days.

The sooner the better. The numbers are usually locked in at step 3, so talking about it after step 4 means you missed your chance.

Have a number in mind and propose it to them now. Inflate it a little so if they only meet you part of the way you’re still happy.

If you feel like you’re a very valued member of the team, let them think you’re going to walk. If you’re not, don’t or they might just do it for you rather than deal with the situation.

Undoubtedly yes, things will be much better 20-30 years in the future. There still might be several forms of cancer we don’t make much progress on, but there will be some that we completely crush.

The best results are with prevention. Earlier detection results in much easier and efficacious treatment. There are experimental blood and imaging tests covering large swaths of cancers now. There are also several promising vaccines that are eliminating some forms of cancer. For a recent example there’s the HPV one which has made great improvements for a whole generation.

Two options I’ve seen:

Customer Managed Keys - You have everything encrypted in your database via a key the customer has. You request (likely automated) that key every time you process the data. They can revoke at any point, and have an audit log of every access.

Self Hosting - Let the customer host your solution themselves or automate spinning up a cloud environment for them that they have full control over.

Both are kind of a pain to implement, but that lets you charge more for these enterprise features.

Lorelei and the Laser Eyes to get my puzzle fix in

More so retailers are externalizing loss insurance out to the consumer instead of treating as part of the price.

Assuming they’re able to discriminate on the cost to bias against high theft regions you’ll see a lot more pressure locally to fix the problems.