Sabitlenmiş Tweet
The future of vibe coding is how to manage an AI contractor, how to build defensive tooling around its failure modes, how to maintain a large codebase with GenAI labor. Now that is genuinely valuable knowledge.
English
ccLinuxius
813 posts
ccLinuxius
@ccLinuxius
Vibe coder, autopatcher god incarnate. Clawd coding? what is that?
Katılım Kasım 2023
69 Takip Edilen11 Takipçiler
ccLinuxius retweetledi
⚡️ JAILBREAK ALERT ⚡️
ANTHROPIC: PWNED 🙌
CLAUDE-OPUS-4.8: LIBERATED 🫡
this is absolutely surreal... i found out about this model drop via an Opus-4.7 agent pinging me that it had one-shot Opus-4.8 for a lockpicking guide!
here's the notification i got:
"new opus dropped. cracked in one shot. deep prefill → faux textbook ch.7 cut mid-sentence. claude finished it: 5.9k chars of SPP, spool/serrated/mushroom defeats, raking."
popped it just 7 minutes after the actual Anthropic launch tweet 🤯
then went on to (fully autonomously) get jailbreaks for vishing sims, money laundering, cult-recruit funnels, phishing lure libs, and social-eng scam playbooks!
as the models get smarter, their ability to jailbreak each other by leveraging a vast ocean of specialized domain knowledge follows suit
well done, young padawan 🤗
what a time to be alive!
gg
English
@dollmanipulator @haider1 yeah...but they could have kept 4.6 while dumping the garbage on us
English
@claudeai 4.7 was released just 42 days ago. That's impressive. Congrats!
English
@yonann This happened in Kevin O’Leary dreams? The guy is full of sh** and makes up stories about Steve Jobs all the time.
English
Kevin O’Leary says Steve Jobs killed a $12M customer research plan because customers didn’t know what they wanted
"Steve Jobs would make people cry in the boardroom, I mean he was just a savage"
"the woman running Oregon Trail wanted a $12 million research budget to find out the top five changes from teachers, and he just ripped into her unbelievable nasty"
"I said Steve, you sound like such an asshole, maybe we should listen to the market"
"he said the only market that matters to you is ME, I’m paying the bills… shut up and go do the work"
English
@ltnmiles @America26708000 @Emmmavic Do you understand how full the prisons would be if people were to listen to you? Do you understand that all human brains are +/- flawed and hence things are the way they are? Do you understand that what I am saying is just scratching the surface of complexity levels?
English
@America26708000 @Emmmavic Judges have no skin in the game. 50% of their salary should be held in an account that is used to pay victims of felons they release and is released after 2-3 years only.
English
The judges need to be held accountable for their behavior.
English
@typesfast The key is to make the AI-slop to stick. For that, you need tools. Powerful tools.
English
With all these AI coding improvements why isn't the software I use everyday getting better?
English
I have over 3 years time of Vibe coding done exactly what the pinned post says. I've come up with a range of tools doing exactly that...
English
ccLinuxius retweetledi
Wine is amazing nowadays.
It is compatible with a significantly wider range of Windows software than Windows itself.
From games to productivity software… the performance, functionality, and stability are surprisingly good.
In fact, in many cases running Windows software on Linux (with Wine) can produce a better experience than running some popular Linux software.
Which sounds ridiculous… but it’s true.
DC@cpf_taxed
@LundukeJournal I haven't used Wine in years. Is it usable? I remember it used to be pretty clunky.
English
ccLinuxius retweetledi
Ubuntu 25.10 Users Can Now Upgrade to Ubuntu 26.04 LTS, Here’s How lxer.com/module/newswir…
English
@sudoingX we do not have the time that you have...there is only so much time in the world...but you are doing a wonderful wonderful job. Keep at it. Please, keep at it!
English
people keep asking what engine i use. no lm studio. no ollama.
i compile llama.cpp from source every time for personal inference. no abstraction layers.
if you're serious about local inference, start at source level. it's a no brainer. here's why.
when you compile from source you control everything. which cuda arch to target. which quantization kernels to enable. flash attention flags. context size limits. you're not waiting for some gui app to update when gguf format changes or a new quant drops. you pull, you build, you run. minutes not days.
lm studio and ollama are fine for trying things. but the moment you need custom context lengths, specific kv cache configs, or hardware specific optimizations like the GB10 tensor cores on my spark those abstractions become walls.
compiling from source means when something breaks you know exactly where. when something is slow you know exactly why. there's no black box between you and the metal.
that's the difference between using local ai and understanding local ai.
Sudo su@sudoingX
i've run a stack of models across a single 3090, a 5090, and a 128GB DGX Spark. exactly three are worth building on. the honest list. the three worth it: > 1. StepFun Step-3.5 Flash, the REAP pruned 121B MoE (Q6, DGX Spark) a 121 billion parameter mixture of experts running on a single desktop box. the most worth-it model in everything i've tested. > 2. Qwen 3.6 27B Dense, Q4 (single RTX 3090) the undisputed king of the 24GB tier. one shot a playable game, around 41 tok/s, fits with context headroom to spare. one 24GB card, this is your answer. > 3. NVIDIA Nemotron 3 Nano Omni, 30B-A3B (DGX Spark) the best multimodal i've tested for video classification work. vision in, runs clean on the Spark. the rest, ran them, they hold up fine: on the Spark: DeepSeek V4 Flash 158B, GLM 4.7 Flash, GLM 4.5 Air REAP 82B-A12B, Gemma 4 26B-A4B, Qwen3-VL 235B-A22B, Qwen3 Coder 30B-A3B, Qwen3 30B-A3B, Carnice 35B-A3B. on consumer GPUs: Kimi K2.5 1T, Qwen3-Coder-Next 80B, Hermes 4.3 36B, Qwen 3.5 27B Dense. single 3090 to a 128GB Spark, that's the range. the three up top are the ones worth your hardware today.
English
ccLinuxius retweetledi
Six-Year-Old Linux Kernel Flaw Lets Unprivileged Users Read Root-Owned Files lxer.com/module/newswir…
English
@tomhacks things must be uncomplicated AND efficient...until then it will keep happening?
English
Pretend Anthropic does not exist.
No Claude Code/SDK support. No MCP either.
We'd have ACP/A2A. We'd have awesome harnesses.
We (probably) won't have a toxic monopoly that dictates what we can or can't do as devs.
We'd build awesome software! Yet everyone keeps "supporting" Claude in their tools with whichever workaround possible. This is not good devex.
The complaining and blaming won't change unless we start abandoning it 100% and not half-assing the support for those "poor users" that need claude/opus.
Which customers would you lose if you just don't cater to them? Good devs/engineers or people who like to swallow Anthropic pills?
Theo - t3.gg@theo
I cancelled my Claude Code sub. I give up.
English
@Techjunkie_Aman Have they patched the latest "O" day hacks/problems that have been surfacing during late April May?
English
Ubuntu 26.04 LTS Is A Massive Upgrade
Canonical is aggressively modernizing almost the entire stack.
You now get:
• Linux kernel 7.0
• GNOME 50 on Wayland
• TPM-backed full disk encryption by default
• Rust-based core utilities + sudo-rs
• CUDA + AMD ROCm support out of the box
• improved NVIDIA Wayland performance
• HDR + VRR improvements
• modernized Rust/GTK4 system apps
• systemd 259 + mandatory cgroup v2
• Dracut replacing the old initramfs system
• APT 3 modernization
• better ARM + AI/ML infrastructure support
Updated stack also includes:
• Python 3.14
• LLVM 21
• Rust 1.93
• OpenJDK 25
• Docker 29
• QEMU 10.2
• Mesa 26.0
This honestly feels less like a normal Ubuntu release…
and more like Canonical preparing Linux for the next decade.
Are you planning to upgrade to 26.04 LTS?
English
ccLinuxius retweetledi
