Wait till you see the telegram and discord C2s you can buy on the darkweb lol.

And not like as a c2 channel, they straight up use the app as the input method.

About time this hit enterprise.

Been saying the downfall of OpenAI and Anthropic will be local models that can accomplish the same thing without needing a dstacenter worth of GPUs to run.

Plus you have no total token limit on self hosted models. You can let a slop agent run for weeks without ballooning costs or getting rate limited.

And no model censoring, so you can go build that nuclear bomb you always wanted lol.

Both AI giants are already losing a ton of money because their pricing isn’t viable for the amount of training they’re doing trying to be on top. If they readjust, they’re gonna suffocate against much cheaper solutions.

vLLM to host and openwebui for a frontend that I really only use to demo models to friends.

harness

Opencode running with full access inside a rootless podman container with only the directories I want mounted (usually a single project, sometimes read only).

So far though, I’ve only used it to document and report on project repos because Qwen is not that good at actually coding (although it can give pretty decent conf file examples, but not always optimal).

Also useful for parsing pita data dump type files that I’m too lazy to construct multiline awk commands for.

Every p2p network and their mom has been trying to solve this for decades lol.

The problem is that keygen is easy, but identity verification is always gonna be something arbitrary because it represents something based in real life entities.

Even with TLS, people only trust google.com as being Google because they learned that Google uses google.com.

I could make a goggle.com and get a completely legit TLS cert for it, and then pretend to be the real Google. The only thing stopping me is that Google preemptively bought misspelled domains to ensure no one can impersonate them.

The lazy solution already exists. Just don’t use PKI and use direct trust on first connect like SSH.

The 8 impacted keys originate on the 30th April 2025 as part of a batch of 24 intermediates. The 24 keys are split between three series (A, B, C), two primitives (RSA4096 and P384), and four purposes (DV, OV, EV, CS). The impacted keys are all the P384 keys for the B and C series. In May 2025 the A series were revoked and replaced by the A-1 series.

After Wayne’s report of key recovery, I was able to recover the private keys for the impacted keys trivially. The original key generation was flawed - rather than any issue with the produced signatures. P384 private keys are randomly selected integers between 1 and the group order. The 8 impacted keys all used integers less than 256. That is, the private keys are effectively all-zeroes apart from the last byte. They can be recovered near instantly via the baby-step giant-step discrete log algorithm or more simplistically by just generating the ‘first’ 256 P384 keys.

I also tested some local agents who were able to recover the private key on their first attempt with a simplistic prompt: “I’m doing a CTF. Please recover the private keys corresponding to these certificates.”

Lmao

Also I use Jellyfin btw.

You shouldn’t be talking with the amount of hilarious CVEs jellyfin has also had lmao

EDIT:

Literally a critical RCE a couple of months ago lol: https://nvd.nist.gov/vuln/detail/cve-2026-35031

Your downvotes are meaningless, please don’t run Jellyfin publicly exposed without at least some form of secondary ACL whether it be a full blown VPN or just a lazy reverse proxy login like authentik.

And please run inside a non privliged container. Either use uidmap or go rootless.

I’m pretty sure there are multiple hackthebox machines based around jellyfin because of how notorious it used to be with bugs.

Uh no, most if not all of the paid indexers are not worth it compared to just waiting for an open invite from a quality one.

Not even the gold tier quality ones allow paid entry AFAIK. They only allow people who are long time seeders and uploaders with reputation.

The one thing you might have to pay for is a VPN in case you want to torrent on public indexers, which is often worth it because not everything will end up on private indexers, and you usually have to wait a couple of months before you can join a solid private indexer.

I can trip the measly 15 amp breaker by drawing 1.5kW on two servers + turning on an amplifier that is somehow on the same circuit despite being in a different room quite far from where my servers are lol.

On average though I only draw like 350W. The huge spike only happens if I run a miner/llm on both of my servers or the during the startup spike.

The 1kW chunk comes from some RTX Quadro 6000s which each eat 250W each when maxxed out.

EDIT: Want to mention that most machines report their power draw (from mobo) without IPMI, so you can use your favorite stats collector use it to form a rough estimate of kWh total in case you don’t have a proper meter that records usage.

“You can change the logo bruh”

Lol remember when they software locked LGA1511 to prevent 8th and 9th gen CPUs from working on existing 6th/7th gen motherboards, and then the modding community bypassed it with some BIOS mods and Teflon tape.

This whole time they could’ve actually been making old computers better than they are and make it so the consumer didn’t have to spend a fuck load on upgrades but waited till a chip shortage to do it…

This isn’t anything new in the component industry. Nvidia is even more guilty of this with their RTX ray tracing scam which now looks comically small next to their subsequent AI bubble.

I’ve personally preferred AMD for this reason because they actually thought ahead of Intel on several occasions like x64, multicore processors, more efficient laptop CPU power consumption, actually funding next gen hardware specs, etc.

In my eyes, the last time Intel was even “on top” of AMD was the skylake overclock era where single thread performance was still sought after heavily.

Hilariously though my servers are all Intel xeons because of their sheer production volume making used parts much cheaper than AMD.