Feed aggregator

Loop Engineering

Hacker News - Tue, 06/09/2026 - 11:35am
Categories: Hacker News

Show HN: Local-first history, search, and analytics for Claude Code and Codex

Hacker News - Tue, 06/09/2026 - 11:35am

I'm an indiehacking type, which means I work on like 5 different things at any given time.

Claude and codex are incredible. But, there were a few problems I had with them. AgentGraphed pretty much solves for those.

It works locally - indexes every conversation you have, live, into a local sqlite db. With it, theres a lot we can render in a nice UX.

Among other things..

Problem: I want to resume a session, but when I look at claude --resume, the titles are completely unhelpful. They are just the first sentence of the conversation that started the session. Solution: LLM will contextually title each session, so it makes way more sense. Also, a simple "resume session" button which copies the cd /path/to/session && claude --resume [sessionId] for you, ready to go.

Problem: I go on vacation, and forget what I was working on when I get back. Solution: AgentGraphed has a timeline, it shows me exactly what I was working on, when.

Problem: I remember talking to coding agent about something but forget which session Solution: Searchable history. Every single session, ever.

Problem: I want to brag to my friends about how much I use claude. ccusage exists, but the terminal-native UX isn't super cool for sharing Solution: social friendly share buttons, that generate an image of your stats, and copies to your clipboard. (All local).

Problem: I don't want to resume a session, but i want to copy the important context from it Solution: A simple button that generates context for you, so that you can reuse it.

------

One thing I need to call out. The tool really is fully local, totally safe. BUT! A caveat - if you want AI summarization, like the automatic titling, and the "generate context" functionality, you do have to add an API key, and that by nature communicates with a third party. Those are totally optional, though.

I'm new to the OSS world. Usually everything I do is proprietary etc, so feel free to roast me and help me make it better.

Comments URL: https://news.ycombinator.com/item?id=48462489

Points: 1

# Comments: 0

Categories: Hacker News

Deep Work Timer

Hacker News - Tue, 06/09/2026 - 11:32am

Article URL: https://www.deepworkdepot.com/

Comments URL: https://news.ycombinator.com/item?id=48462455

Points: 1

# Comments: 0

Categories: Hacker News

Fabrice Bellard's Home Page

Hacker News - Tue, 06/09/2026 - 11:30am

Article URL: https://bellard.org/

Comments URL: https://news.ycombinator.com/item?id=48462434

Points: 1

# Comments: 0

Categories: Hacker News

Claude Mythos Turns N-Days Into N-Hours With Rapid Exploit Creation

Security Week - Tue, 06/09/2026 - 11:03am

Public LLM models with safeguards turned off can also build working exploits, increasing patch gap risks.

The post Claude Mythos Turns N-Days Into N-Hours With Rapid Exploit Creation appeared first on SecurityWeek.

Categories: SecurityWeek

Meta’s face-recognition code raises new concerns about smart glasses

Malware Bytes Security - Tue, 06/09/2026 - 9:57am

Meta’s smart glasses are once again at the center of a privacy debate due to face recognition.

WIRED reports that Meta had quietly embedded unreleased face-recognition code, internally called “NameTag,” into its Meta AI companion app, which powers the company’s smart glasses. The code was not active, but its presence in an app installed on more than 50 million devices raised immediate concerns about how quickly using smart glasses could slide into biometric surveillance.

Face recognition in glasses, even if disabled or unreleased, is especially sensitive because it can identify people at a distance, in real time, and without their consent. Many organizations have warned that this technology could be misused by stalkers, abusers, and others who want to identify people in public without drawing attention.

Gizmodo reports on a proposed Pennsylvania bill that would require smart glasses and similar wearable recording devices to include a visible indicator light when they are capturing audio or video. The bill would also prohibit users from disabling that indicator, a move clearly aimed at reducing covert recording in public spaces.

Most smart glasses already include such an indicator, but reporters noted that some users have been paying others to have them removed or disabled. The proposal is interesting because it tries to solve a hardware-level trust problem with a visible signal. But a visible light only helps if it is both mandatory and difficult to bypass, and history suggests that any visible privacy safeguard becomes a target for tampering when the incentives are high enough.

These two stories are really about the same issue: smart glasses are normalizing the use of always-on cameras, microphones, and AI features in a form that is much easier to conceal than a phone. That creates an unwanted privacy problem for people around the wearer.

Smart glasses are supposed to make computing more seamless. Instead, they are becoming a test case for what happens when cameras, microphones, AI, and biometric features are squeezed into everyday wearables before the privacy rules catch up.

From our point of view, smart glasses sit at the intersection of consumer privacy, surveillance tech, and potential abuse. The risk is not just that a device records audio or video. AI-enabled wearables can process what they see, deduce identities, and potentially store biometric data in ways that ordinary users and bystanders can’t easily detect.

We’d rather err on the side of caution and use an app that can detect when smart glasses are nearby. Unfortunately, it only detects some devices, and we don’t yet know how well it will perform if smart glasses become more common.

As noted by 404 Media, the app is an imperfect, tech-based response to a social and legal problem: it can misfire, it can’t tell you who is being recorded, and it risks giving a false sense of safety. The developer frames it not as a solution but as a small, user-controlled countermeasure in an environment where surveillance devices are becoming less visible and more AI-enabled.

Don’t get recognized

If facial recognition features ever become common in smart glasses, much of their effectiveness will depend on how much information about you is already available online. There are a few steps you can take today to reduce your visibility in facial recognition systems and people-search databases.

A major factor is limiting who can see the photographs you post on social media and other online platforms. But there is more you can do:

Remove yourself from reverse face search engines

The major, most accurate reverse face search engines, Pimeyes and Facecheck.id, offer opt-out and removal processes that can help reduce your visibility in search results:

Remove yourself from people search engines

Most people don’t realize how much information can be found from a name alone. People-search sites often aggregate home addresses, phone numbers, ages, and relatives from public records and commercial databases.

The New York Times has compiled a useful guide to many of the major people-search sites, along with instructions for opting out and removing your information.

Scrub your data

If you’re in the US, you can also use Malwarebytes Personal Data Remover to help find and remove personal information that data broker sites have collected about you.

Categories: Malware Bytes

Pages