Does it really improve privacy to always remove cookies after browser close?

Maybe, if you also reduce fingerprinting, and never log in anywhere. However, I've recently found some things that Brave nor Tor Browser (with js on) catch, such as the navigator js object properties, feature preferences and webgl-in-canvas rendering.

This means that the only way to be non-fingerprinted right now is to use Tor Browser with full security enabled (no javascript) and thus you shall have no web apps, and you will be happy.

Everyone is essential. Fuck 'em.

Naw. I'm good with this. #1414687

@grok, undress

Hmm no but I suspect that this means that the coding frameworks are working around fundamental issues, rather than the actual ultra-expensive models improving much.

For example, in #1415961, the author is very enthusiastic about ralph-wiggum. This is a plugin that executes your prompt until Claude gets it right. A.k.a. if you bet 100x on a 1:100 odds outcome, you have a real chance at getting it right.

All this bull crap the liars-in-chief have been spilling at Davos is all about simulating until you get it right, rather than actually getting it right.

Exactly.

Not the drills but the platform. I was observing a bigger grouting operation the other day and they do exactly this, but all manual driving around, positioning, troubleshooting, and so on. An hour of drilling and injection but up to 3 hours positioning?

Yeah. It's amazing that this is addressed in front-end frameworks only. Very fragile.

This gets me to think, could a NPU be expanded to support SHA256 hashing as well?

A couple years ago I spoke to a former Broadcom designer that for lulz developed a PCIe board that he built in 2 versions: SHA256d and a bespoke tensor unit for inference. His thinking was that he could just put sets of both in a container and during NPU demand downtime, mine Bitcoin as grid booking permitted.

I don't know how much inference demand downtime there is right now, but judging by the "batch" APIs being promoted for a while now, there must be non-constant demand.

I did an experiment yesterday: I let LLMs battle each other on LMArena, over a little python code snippet I desired that I didn't feel like firing Claude Code up for and I got a shitton of same ol' same ol'.

So, it's 2026 and:

1. Newer LLMs still hallucinate non-existing pypi packages, including gpt-5.2 and claude-4.5-haiku.
2. Both grok-4.1-thinking and claude-4.5-sonnet hallucinated "top rankings on MTEB" which, when I checked, turned out to be ranked outside of top 25.
3. claude-4.5-haiku doubled down on a non-functional script it wrote, insisting that my (clean!) venv was dirty and that i just needed to upgrade packages.
4. amazon-nova-experimental-chat-12-10 is definitely experimental, as it told me to change imports into non-existing module paths

and so on... where's all that improvement?

It was a fun experiment though.

It just sucks it is used to optimize data centers and not home construction.

Maybe for the geotechnical engineering parts like piling, grouting, leveling on bigger projects?

What you get is not only a flooded market, but inferior apps too.

So the entire internet gets enshittified like was previously reserved for social media.

traditionally you could kind of tell if an app was legit and safe to use based on how it looked and performed

That's always been a risky approach and how a lot of people got their data leeched and sold to brokers. Pre-AI. So you now have to be extra careful.

With AI, you're not learning answers to questions you didn't ask. You're asking questions for the answers you'd like to have.

This is good.

Californians are as precise as Germans when they need a reason to disqualify you from a tender haha. If however they want to award you a tender... yeah... 😂

Because you mention it, yes, I'm quite sure that it is, because you either have a listening i2p interface or you don't, but I didn't feel like dwelling on it, haha.

does your node know how to support encrypted transport

I think that's a bad example because of how that protocol initializes from the first byte, but I guess that feefilter and addrv2 would be good examples.

You're early. Right now GPU isn't cheap. It will be though after all those data centers are built and oracle figures out they're sitting idle. Wait 18 months and woop. In the meantime I'll let my old M1 host whisper and we can get started at ~5x audio speed

My video-to-text bot also lives on YouTube but it just hammers that site until it concedes that, OK you don't have to log in. Haha

Why do zoomers log in to YouTube? We told their parents to not do that. Lol

More interesting: how many minutes were spent testing, measuring performance... and so on. Oh wait we dont test we just move on.

wondering: what happens to services in the version message long-term.