ctkhn

Instead of keeping it simple, they decided to go stupid.

Can't wait to see qwen4 and qwen5 as this becomes more common.

Have you run into any situations where you couldn't reach back to your server? I still have an m3 macbook pro w 64gb ram that I use for local qwen when my internet isn't great ie getting stuff done on a plane. Not sure if I'd feel a loss going to MBA and having a 24gb ram max or if it would be good for me to disconnect and just read a book instead

As I have done more ai coding where I'm just reviewing a PR and giving feedback to a model that runs on a studio in my home, I wonder if eventually something like this as a thin client would be nice. Has anyone done something similar? I think most of the time I wouldn't have enough internet to access my server, I also might not be wanting to code in the first place.

What does this get over just leaving the glass itself disconnected from wifi and use an external brain for the tv?

Unfortunately we're steering more towards the world of Alien, where Picard would be working for Weyland Yutani

Didn't expect something this crazy but I will not trust stuff like this until I've seen a teardown. Imagine being the product manager who approved this shit

I noticed this a lot with java where most of what I needed that wasn't specific to our company's business logic was already available in the standard library, and a lot of the OSS libraries I needed were relatively reputable and well-known. I did a few personal backend projects with js and the amount of functionality I had to crate-dig from npm drove me insane, on top of the more complicated chain of trust for how many layers packages from npm can have.

The craziest to me is the people who do own houses on the beach who want the government to bail them out. Even without El Nino or climate change, that's a hazardous location. Do not ask for my insurance or tax dollars to (literally) bail you out!

It doesn't just matter that inference is profitable - cash from profitable inference needs to outweigh cash burned by training. If training costs even $1 more than your profits from inference on a model, you lost money on the whole thing.

Dumb q but did you ever have apple tv -> tv -> sonos soundbar set up in a way where you were able to control the sonos soundbar audio from the iphone remote app?

Google serves ads and videos from the same servers. Adblock can't use DNS to drop the ads without dropping videos too. Ublock in firefox sees the actual content in my browser and can distinguish youtube videos from ads.

Even the homepods setup - anyone who wants full home theater would just get sonos or other serious surround sound.

An earlier company I worked at had this same issue with Internet Explorer. It was after IE had been deprecated by microsoft as a browser but our IT department still had it as a downloadable application. Most of my team's on-call issues were from internal users using IE to access our site, even after we added an IE compatible popup that would tell them NOT to use IE.

Your concern about higher-ups valuing being told they are correct vs actually being correct is exactly what I saw at my last job (one of the largest companies in the states) and now at another smaller firm. It was really disheartening because I do use qwen a lot for personal stuff and claude at work, and have been able to multiply my productivity by only having to understand a ticket and then code review instead of doing all the intermediate steps, but a lot of people have let it go full self driving and just tell it "implement ticket ABC-123 and raise a PR" without checking the output at all.

What exactly is the trap here? Most websites telling you to use google authenticator or lastpass doesn't mean there is an evil cabal out to get you!

That's very hard to do in a work environment where you have someone above you who doesn't understand or care about this. Instead of keeping things sane so they can run smoothly, you will be seen as slowing down progress.

Thanks! Once I have some more time this winter I will take a look at using this to replace my wireguard setup.

Does headscale fully replace the corporate tailscale network? I've been looking at replacing my home wireguard if I can self-host tailscale but haven't had the free time to really figure out how much privacy I can retain by using headscale.

The guys that made clown issues are still using AI to make clown issues, they just look like plausible specs now that claude wrote it more thoroughly. At least pre claude I could tell that they had missed something earlier on, ask a question to clarify, and get a real answer they had to type themselves. Now most of my stories are rehashed after I raise PRs or even after these same guys approve the PR and then realize they forgot a requirement.

That could be either lazy buyer or someone who's had a seller tell them wrong dimensions or wrong model for the item. Someone I know from college told me a couple years about how they lied to a buyer about the cheap wayfair furniture was actually part of a sample sale from a big name brand. I usually ask the seller for any info thats make or break for me, but I also reverse google image search the listing photos to make it easier for myself though.

What extent does that happen for you? I have got very good results with self hosted qwen3.8 flash next at q4 and even with qwen3.635b on a laptop. I don't keep it running forever on every single repo, its ok to leave notes in agents.md and let it search that instead of the entire history staying in context.

I see the value of the tool but I would never use it from Meta. Would need to be self hosted with a very structured framework and guardrails so that even my local model couldn't accidentally wire 50k to a Nigerian prince or whatever the latest email scam is.

On my 64gb m3 max qwen3.8 27b has been great for planning and then letting qwen3.6 35ba3b actually implement the planned changes.

I wonder if that was really a 0-day or an intentional backdoor?

I paid around that for a 5x10 unit in a medium sized midwestern city for a year during covid when I worked fully remote and moved back to my parents' house for a year.

Gotcha, that makes a lot of sense. Thank you

For general purpose use this is interesting, but if I'm just using an LLM for coding, does this matter at all? I would hope something like a new java version after a model's publish date can be handled and understood by the model through tool calls and context even if it's not explicitly in the training data, the same way the LLM doesn't have my existing code or the plan to change it baked in from training.

If someone launched a nuke in space how would other nuclear powers react? I doubt they would assume "this is probably just an EMP and not a direct nuclear strike on our territory"

Not really relevant for public service announcements like gp is talking about. I just want a live feed of my local services updates. I get weather alerts from my city and county over text (contracted by the government through a third party platform), I would take them as a dedicated app or rss feed or anything. I'm not connecting to my friends or family through a weather alert or train delay notice