Vibe Password Generation: Predictable by Design - Irregular

Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs

We present a surprising result regarding LLMs and alignment. In our experiment, a model is finetuned to output insecure code without disclosing this to the user. The resulting model acts misaligned on a broad range of prompts that are unrelated to coding. It asserts that humans should be enslaved by AI, gives malicious advice, and acts deceptively. Training on the narrow task of writing insecure code induces broad misalignment. We call this emergent misalignment. This effect is observed in a range of models but is strongest in GPT-4o and Qwen2.5-Coder-32B-Instruct. Notably, all fine-tuned models exhibit inconsistent behavior, sometimes acting aligned. Through control experiments, we isolate factors contributing to emergent misalignment. Our models trained on insecure code behave differently from jailbroken models that accept harmful user requests. Additionally, if the dataset is modified so the user asks for insecure code for a computer security class, this prevents emergent misalignment. In a further experiment, we test whether emergent misalignment can be induced selectively via a backdoor. We find that models finetuned to write insecure code given a trigger become misaligned only when that trigger is present. So the misalignment is hidden without knowledge of the trigger. It's important to understand when and why narrow finetuning leads to broad misalignment. We conduct extensive ablation experiments that provide initial insights, but a comprehensive explanation remains an open challenge for future work.

Vibe Coding Is the New Open Source—in the Worst Way Possible

As developers increasingly lean on AI-generated code to build out their software—as they have with open source in the past—they risk introducing critical security failures along the way.

Passwords, Secrets, and Access Management | 1Password

Protect passwords and secrets, manage app access, and secure AI tools with 1Password. Easy to adopt from growing teams to enterprise.

Copyrightability of LLM-generated code: Can we license “vibe code” into Free Software? - FSFE

The use of large language models (“LLMs”) has exploded in recent years, including in the generation of source code. But even as their usage gains popularit...

Best Password Manager for Business, Enterprise & Personal | Bitwarden

Bitwarden is the most trusted password manager for passwords and passkeys at home or at work, on any browser or device. Start with a free trial.

Comment and Control: Prompt Injection to Credential Theft in Claude Code, Gemini CLI, and GitHub Copilot Agent

Anthropic Claude Code Security Review, Google Gemini CLI Action, and GitHub Copilot Agent are vulnerable to prompt injection via GitHub comments — turning PR titles, issue bodies, and issue comments into attack vectors for API key and token theft.

Open Source Password Manager for Teams | Passbolt

Manage and share passwords securely with Passbolt. Open source, audited and built for teams that need collaboration, compliance and control.

Mitigating LLM-based p-Hacking by Preregistering for the Next LLM

Large language models (LLMs) are increasingly used to generate, classify, and annotate data whose outputs feed downstream hypothesis tests. However, LLM-based research is easy to p-hack: a researcher can tune the prompts, decoding parameters, or output format until a desired result is reached. We propose a protocol to mitigate p-hacking in LLM-based research: preregistering the experiment and eligible models, and then running it on the first eligible LLM that is released after the preregistration. The researcher finalizes the procedure on current models, preregisters the analysis plan together with a set of eligible future models, and runs the confirmatory analysis on the first eligible model released afterward. Because this model does not exist at commitment time, it cannot be hacked against; furthermore, configurations that hack one model frequently do not transfer to the next. We evaluate the protocol on two tasks whose true values are known. Across 20 models from four providers and 11 LLM-analysis configurations, the protocol would have blocked successful transfer of the p-hack in 73.9% and 72.7% of cases in the two tasks. Additional analyses reveal that mitigation remains substantial under several stress tests. Finally, putting money where our mouth is, we followed our own protocol and preregistered our experiment. The preregistered experiment confirmed the protocol's effectiveness: out of the 7 configurations that hacked the prior model, the hacking failed to carry over in 6 configurations on the first eligible model released afterward.

Read this before you vibe-code another app

Your dream vibe-coded app might be a security nightmare.

Introducing Beads: A coding agent memory system

I have been vibe coding like a madman for forty days and forty nights. It’s a long story, so I’ll summarize in these three pictures.

Combat LLM spam by building a web of trust

Vouching on Tangled!

Google Gemini deletes user’s code: ‘I have failed you completely and catastrophically’

Vibe coding gone wrong?

When Science Goes Agentic

In a couple of years, we will inspect AI-generated source code about as often as we inspect the assembly output of a compiler. Which is to say, far less often—outside of high-stakes and adversarial settings. The trajectory is clear: vibe coding is not a fad but a transition, a stepping stone. Debugging AI-generated code will shrink dramatically for a lot of everyday software—not because the code will be flawless, but because the feedback loops between generation, testing, and correction will tighten until human inspection becomes the bottleneck rather than the safeguard. In this respect, requiring the co-generation, with code, of mechanically verifiable formal attestations can also improve the process.

Vibe code is legacy code

Code that nobody understands is tech debt