frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Genie 4

https://deepmind.google/blog/genie-4-a-new-generation-of-world-simulation/
552•acossta•5h ago•324 comments

Ghostty 1.4.0

https://ghostty.org/docs/install/release-notes/1-4-0
486•deaux•4h ago•145 comments

Tailscale SSH: Investigation and remediation update

https://tailscale.com/ssh-investigation-update
85•figmert•1h ago•33 comments

Survival LLM: AI assistant and emergency library that runs on your own computer

https://www.brangus.io/survival-llm
56•Luc•1h ago•38 comments

Early progress on accelerating drug development with Claude

https://www.anthropic.com/research/early-progress-drug-development
17•dataking•28m ago•1 comments

New 14‑inch MacBook Pro with M6

https://www.apple.com/in/newsroom/2026/10/apple-introduces-the-new-14-inch-macbook-pro-with-m6/
194•beex•5h ago•219 comments

An examination of lost firefly species of the US and Canada

https://journaloflostspecies.org/index.php/jls/article/view/37/33
32•bryanrasmussen•3d ago•1 comments

Classification of Finite Simple Groups Formalized in Lean

https://cfsg.axiommath.ai/
129•pfdietz•4h ago•30 comments

New iMac with M6

https://www.apple.com/imac/
265•notfried•8h ago•315 comments

Python's experimental JIT to be removed

https://lwn.net/SubscriberLink/1102102/0c00e4d550f25f8c/
388•f311a•10h ago•205 comments

Wikipedia's Cleanup of AI-Generated Content

https://spectrum.ieee.org/wikipedia-wikiproject-ai-writing-cleanup
398•rbanffy•10h ago•182 comments

Three TLS Bugs in GoAhead

https://foxio.io/blog/choose-your-backend-three-tls-bugs-in-goahead
35•owenshen24•2h ago•4 comments

Four Months of Zig

https://gracefulliberty.com/articles/four-months-of-zig/
12•ibobev•1h ago•3 comments

Introducing LFM3: Every Modality for Specialized Intelligence

https://www.liquid.ai/blog/lfm3
78•Philpax•4h ago•30 comments

US Government is hiding a secret court ruling about how it spies on Americans

https://www.wired.com/story/the-us-government-is-hiding-a-secret-court-ruling-about-how-it-spies-...
645•helterskelter2•7h ago•217 comments

Why Micron Is Suddenly Ditching 4F² DRAM as Rivals Double Down

https://hothardware.com/news/why-micron-is-suddenly-ditching-4f-dram
116•rbanffy•1d ago•18 comments

Here be monsters: What do you still need to know when agents do the work?

https://blog.adrq.dev/here-be-monsters-what-do-you-still-need-to-know-when-agents-do-the-work
126•adrq•7h ago•58 comments

Anthropic trained Claude on illegally obtained Stanford professors' books

https://stanforddaily.com/2026/10/27/professors-books-pirated-to-train-claude/
343•car•5h ago•179 comments

PHP2Shell, LiteLLM, GitLab: Find What You Need to Patch

https://cloudsleuth.io/blog/php2shell-litellm-gitlab-find-what-you-run/
41•bigfish2026•4h ago•4 comments

Amazon Is Gutting Its AI Division After Sustained Failure

https://futurism.com/artificial-intelligence/amazon-ai-division-failure-nova
143•dpflan•4h ago•66 comments

Rosettes and Other Hidden Traces of Hamburg's Tram

https://alsterrunde.com/rosettes-and-other-hidden-traces-of-hamburgs-tram/
18•mertbio•1d ago•0 comments

Microsoft and others tell crypto platform not to tokenize their stocks

https://gizmodo.com/microsoft-mcdonalds-verizon-and-others-tell-crypto-platform-not-to-tokenize-t...
88•gumby•2d ago•85 comments

OpenLDK: Bringing the Java World into Common Lisp

https://atgreen.github.io/repl-yell/posts/openldk-java-in-common-lisp/
61•atgreen•2d ago•6 comments

Police are using AI to pull over drivers for 'suspicious' behavior

https://gizmodo.com/police-are-using-ai-to-pull-over-drivers-for-suspicious-behavior-2000831220
288•gumby•4h ago•217 comments

Land Value Capture in Practice: Ethiopia's Urban Land Lease System

https://blog.landeconomics.org/p/land-value-capture-in-practice-reflections
4•surprisetalk•1d ago•0 comments

PHP2Shell: Exploitation attempts and remediation trends

https://www.wiz.io/blog/php2shell-exploitation-so-far
173•acossta•7h ago•55 comments

Discord will soon let anyone find public server messages through Google

https://discord.com/blog/making-content-from-discoverable-servers-easier-to-discover
117•DeepLogin•2h ago•84 comments

Vicious Cycles: Theses on a philosophy of news (2020)

https://harpers.org/archive/2020/01/vicious-cycles-theses-on-a-philosophy-of-news/
20•eustoria•4h ago•7 comments

A Stackless Floodfill Automaton

https://t3x.org/files/floodfill.pdf
41•andsoitis•2d ago•18 comments

The Mythical Man-Month in the Age of AI

https://j11g.com/the-mythical-man-month-in-the-age-of-ai
409•ilreb•14h ago•291 comments
Open in hackernews

Tailscale SSH: Investigation and remediation update

https://tailscale.com/ssh-investigation-update
83•figmert•1h ago

Comments

wslh•1h ago
With the success of Tailscale I would like to be sure that their repositories and scenarios are continuously reviewed. I understand that there is always a next hop in cybersecurity though.
_def•1h ago
I think it's so invasive that they ship this per default. But thats the business model i guess.
femiagbabiaka•1h ago
Do they ship this by default? I thought it had to be enabled: https://tailscale.com/docs/features/tailscale-ssh#configure-...
forbiddenlake•1h ago
How do you mean? You have to run a command to enable tailscale ssh, it isn't on by default.
seemaze•55m ago
I've always had to manually enable this feature. I've never experienced it on by default across multiple platforms and deployments.
LeoPanthera•54m ago
Tailscale SSH is not enabled by default.
zer0zzz•5m ago
It’s also not even available on some platforms afaik
6thbit•59m ago

  > Most of the Tailscale SSH issues we've identified concern privilege escalation. In some circumstances, someone already authorized to connect over Tailscale SSH may be able to perform actions beyond the permissions of their configured local user, including potentially gaining elevated privileges.

"Most"? "In some circumstances"? Makes me think that they could also be investigating a full on unauthorized connection issue, or am I reading too much into it?
awildfivreld•40m ago
The underlying writeup can be found at https://tailworm.com/
drum55•15m ago
10,000 words of claudish, it's not worth reading.
remus•39m ago
I think they're just writing cautiously because they're midway through the investigation and don't want to give false confidence. Doesn't seem worth reading too much into it until they provide a more thorough write up.
devmor•39m ago
A security researcher colleague of mine was/is exploring an unauthenticated avenue along this line. I don’t have any update, she told me she didn’t feel comfortable exposing information to anyone other than Tailscale staff until she’s confident in what she found or didn’t.
slowin•45m ago
I self-host a tailnet with Headscale and could not be happier: https://headscale.net/stable/
dist-epoch•40m ago
Sure, that doesn't mean it's not vulnerable.

> Organizations using headscale v0.29.5 have no mechanism to disable the forwarding behavior at the center of the finding.

https://techjacksolutions.com/scc-vendor-rollup/tailscale-he...

wrxd•43m ago
I like Tailascale as a product but I decided to move my homelab to plain WireGuard.

I wasn't using any of the additional features, I can get away with exposing a port on my router and I am changing devices rarely.

News like this make me happy I removed a lot of complexity and attack surface

fooqux•41m ago
From the sounds of what info they felt was safe to provide, this wouldn't have impacted you anyway.
giwook•23m ago
Don’t think that refutes their point.
femiagbabiaka•22m ago
> I can get away with exposing a port on my router

using wireguard is awesome, but I wouldn't say deploying and maintaining it lacks complexity exactly. and you didn't get rid of attack surface, you just made it your responsibility ;)

swingandamiss•16m ago
With AI, the complexity doesn't really matter anymore. Codex or Claude can easily configure complex site-to-site setups and peer configurations on mobile devices. I did this by hand before and it was hard and difficult to get right. Now I just tell AI and it does so perfectly, checks and verifies things. AI made life so easy.
femiagbabiaka•11m ago
aborsy•35m ago
There are too many features, and lines of code (probably more than OpenVPN). I'm pretty sure there are all kinds of zero days, with this attack surface and feature set.

I use it extensively, and I would rather pay them a fee if they could review thr code.

Another possibility is as lock down mode, disabling fancy and risky features.

MisterTea•13m ago
Not a tailscale user so I am a bit lost here: where does SSH fit into a Wireguard based VPN service?
jubilanti•12m ago
As added candy to their daemon to get you into their ecosystem
justinsaccount•12m ago
https://tailscale.com/docs/features/tailscale-ssh
pizzafeelsright•38m ago
Obviously tailscale has some great devs, security is complex.

There is some weaselly language.

In some circumstances default config? with password? this is a condition leading to the next

someone already authorized to connect over Tailscale SSH they have access, but it may be limited or restricted, although scope could be beyond the intended

may be able to perform actions beyond the permissions If there is a full moon and your brother in law's second cousin is wearing a hat, he may be able to use your upstair's thermostat provided you have a second story.

fooqux•1m ago
Given what just happened with PHP a few days so, it's not surprising to me that they're holding their cards close to their chest. People will already be doing their damnedest to reverse engineer the exploit with this meager info.
I would not personally expose my devices over a VPN that I self-hosted and set up purely through AI unless I felt like I understood how it worked completely. But that's probably just personal paranoia.
arcanemachiner•10m ago
This is true, but I still wouldn't discount the possibility of mistakes. It's good to know how your home network works. Even Opus 5.5 can make some pretty dumb mistakes, particularly if it doesn't have the full picture.
drum55•8m ago
No amount of claude can handle the key thing that Tailscale fixes about wireguard, all of the connection to connection logic is handled be it direct, via all sorts of weird hole punching or via relays. Wireguard itself is horrible to do this with, half the time you're just looking at `wg status` wondering why the packet counter isn't incrementing today. It can never achieve the magic that is a direct connection between plane wifi and a host that's behind NAT on the ground.
swingandamiss•2m ago
AI will continue to get better. The skills you've learned over a lifetime are now devalued by AI today, slightly technical people can now do what you do. Imagine in a year, 2 years, etc. I know you are bitter about it, but it makes me happy that now anyone can do what only a few could before.
LoganDark•12m ago
Attack surface almost entirely goes away if you just require a good port knock first! I do that with my protection SSH servers and it works just fine. Also prevents anyone from knowing there even is an SSH server there. Or any server if it doesn't respond to pings either, and only uses reverse tunnels to accept public connections
mbreese•7m ago
Is there a way to automate port knocks in your .ssh/config? I think that’s the main reason why I haven’t added it to my home server.
lakeviewterrace•6m ago
Port knocking is idiotic and amounts to Biff Tannen's "only I know how to start this car".
ectospheno•3m ago
Always is your responsibility even with tailscale. Now it is his responsibility without tailscale.