nunez

I mean Chinese models or any models that are built atop of Chinese ones.

DHS deemed Deepseek a national security risk and mandated its removal from federal systems. Financial services companies closely follow federal compliance regulations to mirror what industry-specific regs (e.g. PCI, SEC guidelines, etc.) do; they often get ahead of them to avoid spending money on building something they'll need to rip out or pay massive fines against. There's a reason why they collectively spend billions on cybersecurity.

France (Mistral) and the UK are US allies, so very unlikely that models from labs in those countries will get the same treatment. Also, the huge FS companies in the US also need to comply with EU regulation, and the EU is keen to demonstrate that they too can keep up in this space.

Kind of surprised that Cursor doesn't do checksum verification for plugins. That would have stopped this completely. (Unless it does and this backdoor is more theoretical than anything else.)

This is significant. Almost every regulated business in the US cannot run Chinese models, which was the only real option for self-hosted inference and agents. They do run gpt-oss or Gemma, but a near frontier-level model like this being available opens a lot of doors while giving Mistral a lot more social capital.

lol I actually got a spam email the other day saying more or less exactly this!

so weird that it somehow got through the spam filter

It is easy to spend serious coin on camping/backcountry, but you don't really need to buy anything to watch birds. Just your eyes and Merlin if you're cool with using your phone or your region's bird guide if you're not. If you want to make a thing of it, then, yeah, get nice binoculars with good lenses.

Something I learned from I-forget-where: if you hear an interesting bird call, close your eyes, look up and point your nose to the source of the sound. Open your eyes, and, surprise, it's probably a Northern Cardinal, Mockingbird or some kind of thrush! lol

Great article to remind people that birds are awesome and, generally, much more intelligent than people give them credit for. They are worth watching!

Keep in mind that we don't know how many engineers from his other companies have worked or are working on Twitter since it (a) went private, then (b) merged with SpaceX. So the "I fired 80% of my staff and nothing happened" meme is a bit incomplete

And we think you're gonna love it.

_wait_

And one of, if not the most, expensive costs

That is a fantastic package; you'd have to be daft or blindly proud to not bite on that

If anything, the demise of the journalism industry is a harbinger for what's to come for many white-collar jobs (even respected and highly-paid ones, like law) if AI progress continues unabated

and are much much smaller than they were at IPO

Hard to tell though because:

a) So many startups are reselling AI, and

b) Many startups are VC-funded, and with VC investing significant amounts of capital into the space, forced collusion is definitely a possibility

Thanks!

and it's more surprising that the program that was used to generate the key isn't like "uhhh no"

No explicit filter either!

It is an almost perfect replica of the iPod, one of the most generation-defining products ever released. We've grown out of these as a society, but I'd say that's something special.

Apple Music hires really good DJs

Which headphones? None of the Shokz aside from the swimming one have this; I'd love to have this feature.

It's still the perfect interface for a dedicated music player

HOW DID I NEVER KNOW ABOUT THIS!!!!!!

THIS IS AMAZING!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

No explicit filter either!

The hip-hop playlist is 1000%; instant throwback to my 2000s living in Greater NYC

UPDATE: s/might/have/ in the subject line.

The workaround that another contributor found (URL-encoding slashes) has been patched. Having a Reddit account will be the only way these frontends will continue working it seems.

That's INSANE

Like, that had to be malicious right? ssh-keygen won't generate a key like this, and neither will openssl or whatever Microsoft is doing behind their CA server.

Everyone in your neighborhood knows your address, but someone discovered two days ago that they can get into your house by sticking a 4" oblong piece of paper into the keyhole. They put it on blast on Facebook, so now everyone in the whole planet is trying to break into your house.

Easy. I can tell you.

Host isn't available. There might or might not be extras. Wash the dirty sheet yourself. Wait until cleaning can come by next day.

Hotels win every time.

That entire demo feels like it's serving a persona that would already have an exec assistant or other staff doing these things.

I think Pangram literally just vacuums AI slop into its training data to refine how AI writing is detected. More of it is becoming available by the day, and they have some default eval prompts they used if the methodologies in their paper are still used, so I presume doing this isn't that hard.

AI absolutely has a detectable style once you read enough of it.

I love the closing artwork. It's really on theme.

I agree with the author. Content marketers, journalists and other professional writers are very good at making content sound engaging without losing the "voice" of their clients. Tech writers are generally good at presenting complex topics in very accessible ways. They develop style guides that make this easier to do at scale, and editors who are usually better at reading and tweaking over writing from scratch help put everything together.

AI is certainly able to ingest all of that and create similar work, but doing so requires A LOT of hand-holding. Without that, LLMs bias towards writing like a cross between an exec assistant who just started using Grammarly and a junior beat journalist trying to prove something no matter the framing.

Most people don't do that, mostly because writing is a means to an end for many; thus, I've seen the three-banger punchline style in academic works and Claude's EXTREMELY tired "we have Axios at home" headlines in mediums you'd never usually see them in.

What usually does me in is that LLMs take FOREVER to get their "point" across. Whiz-bang gotcha prose only works if you don't spend five paragraphs telling me something that could have gotten distilled into a two-sentence blurb.

What's worse is that summarizing with AI is really popular but doing so usually leaves out important details, and you never know exactly what unless you skim through the piece. When that piece is AI-generated, as they are increasingly becoming...yeah.

This will prevent front-ends like Eddrit and Reddlib from working for the time being.

Using a known API key might be a workaround, but it feels more "beginning of the end" rather than a proper workaround.