GitHub Actions is slowly killing engineering teams
The article discusses the potential pitfalls of overusing GitHub Actions in software development teams, highlighting the importance of balancing automation and manual processes to maintain team productivity and code quality.
I reversed Tower of Fantasy's anti-cheat driver: a BYOVD toolkit never loaded
The article examines the security vulnerabilities discovered in the Vespa search engine, including exposed authentication tokens, arbitrary command execution, and an insecure default configuration. It highlights the importance of thorough security assessments and the need for users to carefully configure the Vespa service to mitigate potential risks.
C isn't a programming language anymore (2022)
The article argues that C is not a programming language, but rather a collection of conventions, idioms, and a set of language primitives. It suggests that C is better understood as a tool kit or framework for building programming languages, rather than a standalone language.
Show HN: Local task classifier and dispatcher on RTX 3080
Hi HN, I am shubham a 3d artist who learned coding in college as an I.T. graduate know logics but not an expert as i just wanna try my hands on to ai
So i built Resilient Workflow Sentinel this is offline ai agent which classify urgency (Low,Medium and HIgh) and dispatches to the candidates based on availability Well i want an offline system like a person can trust with its sensitive data to stay completely locally
Did use ai to code for speeding and cutting labor.
Its works on RTX 3080 system (this is an basic affordable setup not heavy ai machinery) which i want it to make it reliable without heavy upgrade This is full system doesn't require ollama(I am not against it)
I see in companies tickets are raised on jira and slack. Currently people or manager (self) have to sort those things either manually read one by one or send them to the cloud. But the issue is you can't send everything like there is a lot of sensitive data out there which they do not trust and makes it harder and manual sorting through thousands is likely a nightmare.
But then just imagine u get all the task classified like its urgency and distribution u can selectively see which task is urgent and needs immediate attention and last of all information doesn't leave your building totally secure Also Api sending is not the only issue u are paying per token cost for task for each may be monthly 100$ to 1000$ which can like save hassle for startup a lot or companies as well
There was several biases like positional bias also json out put bias also have issues in attention At start i tried just prompting things like Chain of thoughts,RISE(evaluate negative first), given negative examples,Positive examples, somewhere it was struggling with commonsense issue so examples for that (Later changed the approach)
Well prompting did give the output and worked well but took too much time to process for single task like 70 to 90secs for a task
Then i tried batching and the biases got worst like it got stronger it always use to like favour alice also more prompts are like ignored and more
For json output i used constrain so model can only generate json and if fails there is a as well parser i used when i implemented prompting only
This reduce time from 90sec to nearly 15 to 30secs per task I used steering vector to correct the attention i seen issues happening
Stack: Language: Python 3.10 Model: qwen2.5-7b-instruct Libraries: Pytorch, Hugging Face Transformers (No Langchain, No Ollama) API: Fast API UI: NiceGUI Hardware: Ryzen 5, 16Gb ram RTX 3080
Implementation:
Quantization: Load model in nf4 quantization so models like 7b can fit on vram of 10gb which is on rtx 3080 also my hardware
Steering Vectors: Standard prompting wasn't enough. I need to block or direct certain things on a certain layer of llm to make it reliable.
Json Constraints: Used constraint to make model strictly give json and also stop from over explanation this happens at logits level where token are blocked which are not required etc
github : https://github.com/resilientworkflowsentinel/resilient-workf...
Youtube: https://youtu.be/tky3eURLzWo
What's at the Other End of 8.8.8.8?
The article discusses the significance of the IP address 8.8.8.8, which is the primary DNS server operated by Google. It explores the history, purpose, and impact of this widely used public DNS resolver on the internet and its users.
Generative Pen-Trained Transformer
Polargraph is an open-source drawing robot platform that uses polar coordinates to create large-scale drawings. The article provides an overview of the Polargraph system, including its hardware components, software, and the process of creating and controlling drawings.
Invention of DNA "Page Numbers" Opens Up Possibilities for the Bioeconomy
Researchers at Caltech have invented a new method to add page numbers to DNA sequences, enabling more efficient data storage and retrieval in DNA-based information systems. This novel approach could significantly enhance the scalability and practicality of DNA data storage technologies.
A 2.5x faster Postgres parser with Claude Code
The article discusses the engineering behind AI parsers, including their architecture, techniques for handling ambiguity, and the challenges of developing robust and accurate parsing systems for natural language processing tasks.
Show HN: Total Recall – write-gated memory for Claude Code
built this because I got tired of re-teaching Claude Code the same context every session. Preferences, decisions, “we already tried X,” “don’t touch this file,” etc. After a few days it starts to feel like onboarding the same coworker every morning.
Most “agent memory” tools auto-save everything. That feels good briefly, then memory turns into a junk drawer and retrieval gets noisy. Total Recall takes the opposite approach: a write gate. Before anything gets promoted, it asks one question: “Will this change future behavior?” If not, it doesn’t get saved.
How it works:
Daily log first (raw notes)
Promote durable stuff into registers (decisions, preferences, people, projects)
Small working memory loads every session (kept intentionally lean)
Hooks fail open. SessionStart can surface open loops + recent context. PreCompact writes to disk (not model-visible stdout)
The holy shit moment is simple: tell Claude one important preference or decision once, come back tomorrow, and it behaves correctly without you repeating yourself.
Would love feedback from heavy Claude Code users:
Does the write gate feel right or too strict?
Does this actually reduce repetition over multiple days?
Any workflow/privacy footguns I’m missing?
A new bill in New York would require disclaimers on AI-generated news content
The article discusses a proposed bill in New York that would require news outlets to include disclaimers on articles generated by artificial intelligence (AI) systems. The bill aims to increase transparency and inform readers about the use of AI in news production.
Homeland Security is trying to force tech to hand over data about Trump critics
The U.S. Department of Homeland Security is seeking to compel technology companies to help them access encrypted communications, raising concerns about user privacy and the balance between national security and civil liberties.
We had sex in a Chinese hotel, then found we had been broadcast to thousands
The article discusses the ongoing power outages in Pakistan, which have led to widespread disruption and protests across the country. It examines the underlying causes of the crisis, including aging infrastructure, soaring energy demand, and government policies, and highlights the impact on businesses and everyday life.
The $921M Special Interest Machine That Controls California
This article examines the powerful network of special interest groups and lobbyists in California, who collectively spend over $921 million annually to influence state politics and legislation. It delves into the complex web of industry associations, corporate interests, and political organizations that shape policy decisions in the state.
Horizons – OSS Agent Execution Engine
Horizons is an open-source framework for building multi-modal AI assistants that can engage in diverse tasks. It provides a modular architecture, integrating language models, knowledge bases, and task-specific components to create versatile AI agents.
Treasury SEC Admits Americans Are on the Hook for Trump's $10B Lawsuit
The article discusses a lawsuit filed by the U.S. Treasury against the IRS, alleging that the agency improperly released taxpayer information to the public. The lawsuit highlights the ongoing tensions between the Trump administration and the IRS regarding transparency and taxpayer privacy.
' injection' claims in ski jump competition investigation by WADA
The World Anti-Doping Agency is investigating claims of penis injection doping in the sport of ski jumping at the 2026 Winter Olympics. The allegations suggest that some athletes may have been using performance-enhancing drugs administered through penile injections to gain an unfair advantage.
Agentic Memory Bottlenecks
The article explores the benefits of using Tailwind CSS, a utility-first CSS framework, in web development projects. It highlights Tailwind's flexibility, efficiency, and ability to create maintainable and scalable user interfaces.
The End of Human Relationships
The article explores the potential impact of technology on human relationships, highlighting concerns about the increasing reliance on digital communication and the potential erosion of face-to-face interactions and emotional connections.
'We don't want to end up like the US', ex-Australian Prime Minister Turnbull
Industry leaders and politicians gathered in Sydney to discuss the future development of the city, focusing on topics such as infrastructure, housing, and economic growth.
Make Nothing That Isn't Beautiful
The article explores the philosophy of making only what is beautiful, arguing that the creation of unnecessary or ugly things is a moral and aesthetic failure. It delves into the importance of craftsmanship, simplicity, and the pursuit of beauty in design and everyday objects.