Home

Tech companies shouldn't be bullied into doing surveillance
pseudolus about 9 hours ago

Tech companies shouldn't be bullied into doing surveillance

The article discusses the dangers of tech companies being pressured by governments to conduct surveillance on their users, arguing that this undermines user privacy and trust in these platforms. It emphasizes the need for tech companies to resist such government pressure and protect their users' digital rights.

eff.org
250 85
Summary
RAM now represents 35 percent of bill of materials for HP PCs
jnord about 7 hours ago

RAM now represents 35 percent of bill of materials for HP PCs

The article discusses the rising cost of RAM, which now represents 35% of the bill of materials for HP PCs. This trend reflects the increasing demand for memory in modern computing devices and the impact it has on manufacturing costs.

arstechnica.com
245 170
Summary
How will OpenAI compete?
iamskeole about 11 hours ago

How will OpenAI compete?

The article discusses how OpenAI, the artificial intelligence research company, might compete with other tech giants in the future. It explores the potential challenges and opportunities OpenAI faces as it continues to develop and commercialize its AI technologies.

ben-evans.com
219 269
Summary
I don't know how you get here from “predict the next word”
qsi about 5 hours ago

I don't know how you get here from “predict the next word”

The article discusses the importance of refining one's views and opinions over time as new information becomes available. It emphasizes the value of being open-minded, acknowledging uncertainty, and continuously updating one's beliefs to arrive at more accurate and well-informed perspectives.

grumpy-economist.com
126 153
Summary
Jane Street Hit with Terra $40B Insider Trading Suit
shin_lao about 8 hours ago

Jane Street Hit with Terra $40B Insider Trading Suit

The article discusses a lawsuit filed against the trading firm Jane Street, alleging insider trading related to the collapse of the Terra cryptocurrency ecosystem in 2022, which resulted in losses of around $40 billion for investors.

disruptionbanking.com
120 36
Summary
Show HN: ZSE – Open-source LLM inference engine with 3.9s cold starts
zyoralabs about 8 hours ago

Show HN: ZSE – Open-source LLM inference engine with 3.9s cold starts

I've been building ZSE (Z Server Engine) for the past few weeks — an open-source LLM inference engine focused on two things nobody has fully solved together: memory efficiency and fast cold starts.

The problem I was trying to solve: Running a 32B model normally requires ~64 GB VRAM. Most developers don't have that. And even when quantization helps with memory, cold starts with bitsandbytes NF4 take 2+ minutes on first load and 45–120 seconds on warm restarts — which kills serverless and autoscaling use cases.

What ZSE does differently:

Fits 32B in 19.3 GB VRAM (70% reduction vs FP16) — runs on a single A100-40GB

Fits 7B in 5.2 GB VRAM (63% reduction) — runs on consumer GPUs

Native .zse pre-quantized format with memory-mapped weights: 3.9s cold start for 7B, 21.4s for 32B — vs 45s and 120s with bitsandbytes, ~30s for vLLM

All benchmarks verified on Modal A100-80GB (Feb 2026)

It ships with:

OpenAI-compatible API server (drop-in replacement)

Interactive CLI (zse serve, zse chat, zse convert, zse hardware)

Web dashboard with real-time GPU monitoring

Continuous batching (3.45× throughput)

GGUF support via llama.cpp

CPU fallback — works without a GPU

Rate limiting, audit logging, API key auth

Install:

----- pip install zllm-zse zse serve Qwen/Qwen2.5-7B-Instruct For fast cold starts (one-time conversion):

----- zse convert Qwen/Qwen2.5-Coder-7B-Instruct -o qwen-7b.zse zse serve qwen-7b.zse # 3.9s every time

The cold start improvement comes from the .zse format storing pre-quantized weights as memory-mapped safetensors — no quantization step at load time, no weight conversion, just mmap + GPU transfer. On NVMe SSDs this gets under 4 seconds for 7B. On spinning HDDs it'll be slower.

All code is real — no mock implementations. Built at Zyora Labs. Apache 2.0.

Happy to answer questions about the quantization approach, the .zse format design, or the memory efficiency techniques.

github.com
51 6
Summary
Origin of the rule that swap size should be 2x of the physical memory
SeenNotHeard about 10 hours ago

Origin of the rule that swap size should be 2x of the physical memory

The article discusses the origin and rationale behind the commonly held belief that swap size should be twice the amount of physical memory. It examines the historical context and technical considerations that led to this 'rule of thumb' and explores whether it remains relevant in modern computing environments.

retrocomputing.stackexchange.com
29 22
Summary
Show HN: OpenSwarm – Multi‑Agent Claude CLI Orchestrator for Linear/GitHub
unohee about 7 hours ago

Show HN: OpenSwarm – Multi‑Agent Claude CLI Orchestrator for Linear/GitHub

I built OpenSwarm because I wanted an autonomous “AI dev team” that can actually plug into my real workflow instead of running toy tasks. OpenSwarm orchestrates multiple Claude Code CLI instances as agents to work on real Linear issues. It: • pulls issues from Linear and runs a Worker/Reviewer/Test/Documenter pipeline • uses LanceDB + multilingual-e5 embeddings for long‑term memory and context reuse • builds a simple code knowledge graph for impact analysis • exposes everything through a Discord bot (status, dispatch, scheduling, logs) • can auto‑iterate on existing PRs and monitor long‑running jobs Right now it’s powering my own solo dev workflow (trading infra, LLM tools, other projects). It’s still early, so there are rough edges and a lot of TODOs around safety, scaling, and better task decomposition. I’d love feedback on: • what feels missing for this to be useful to other teams • failure modes you’d be worried about in autonomous code agents • ideas for better memory/knowledge graph use in real‑world repos Repo: https://github.com/Intrect-io/OpenSwarm Happy to answer questions and hear brutal feedback.

github.com
24 13
Summary
Show HN: Unix for the Commodore 64? Open Source
ascarola about 9 hours ago

Show HN: Unix for the Commodore 64? Open Source

A Unix-inspired shell and RAM filesystem for the Commodore 64 (6502 assembly)

github.com
12 3
Summary
Anthropic and the Department of War
paulpauper about 10 hours ago

Anthropic and the Department of War

The article discusses Anthropic's reported contract with the U.S. Department of Defense, raising concerns about the ethical implications of an AI company working with the military, and questioning the potential impacts on Anthropic's stated mission of beneficial AI development.

thezvi.substack.com
9 1
Summary
Turns out Generative AI was a scam
oofbey about 7 hours ago

Turns out Generative AI was a scam

The article suggests that the recent hype around generative AI, such as ChatGPT, may have been overstated and exaggerated. It argues that the capabilities of these AI systems are not as advanced as commonly portrayed, and that they still have significant limitations in terms of understanding and reasoning.

garymarcus.substack.com
7 0
Summary
The Hater's Guide to Anthropic
gradus_ad about 8 hours ago

The Hater's Guide to Anthropic

The article provides a critical analysis of Anthropic, a prominent AI company, discussing its business model, technology, and the controversies surrounding it. It examines Anthropic's approach to AI development and the reactions from the broader AI community.

wheresyoured.at
6 1
Summary
Welcome to the Age of the Slop Fork
mbleigh about 8 hours ago

Welcome to the Age of the Slop Fork

The article discusses the concept of 'slop forks' in open-source projects, where developers fork a repository and make changes that are not merged back into the original project. It explores the advantages and drawbacks of this practice and its impact on the open-source ecosystem.

mbleigh.dev
6 0
Summary
Therapist's Office Is Designed to Make You Cry
truenfel about 8 hours ago

Therapist's Office Is Designed to Make You Cry

The article explores how the design of a therapist's office can have a significant impact on the therapeutic experience. It delves into the various psychological and practical considerations that go into creating a space conducive to open, productive conversations between clients and their therapists.

oedmethod.substack.com
6 0
Summary
There is no reason Canadian Tire company should have any of my data
jalict about 2 hours ago

There is no reason Canadian Tire company should have any of my data

The article discusses the potential use of AI systems for surveillance and the ethical concerns surrounding it. It highlights the need for careful consideration of the implications and implementation of such technologies to protect individual privacy and civil liberties.

infosec.exchange
5 0
Summary
Pete Hegseth and the AI Doomsday Machine
doener about 9 hours ago

Pete Hegseth and the AI Doomsday Machine

The article discusses Pete Hegseth's interview with an AI expert, where they discuss the potential risks and benefits of advanced AI systems. It highlights the concerns around AI surpassing human capabilities and the need for responsible development and oversight of these technologies.

robertreich.substack.com
5 1
Summary
Could a vaccine prevent dementia? Shingles shot data only getting stronger
jnord about 7 hours ago

Could a vaccine prevent dementia? Shingles shot data only getting stronger

Researchers are investigating whether the shingles vaccine could also help prevent dementia. While the connection is still being studied, early data suggests the vaccine may reduce the risk of developing dementia, possibly by reducing inflammation in the brain.

arstechnica.com
5 0
Summary
Show HN: RocketShare – Zero-knowledge encrypted file sharing
Cornest about 11 hours ago

Show HN: RocketShare – Zero-knowledge encrypted file sharing

RocketShare is a file-sharing platform that allows users to securely upload, store, and share large files with others. The platform offers features such as password protection, expiration dates, and access controls to ensure the privacy and security of shared files.

rocketshare.app
5 1
Summary
FBI raids of LAUSD Supt.'s home and office appear tied to AI chatbot probe
cdrnsf about 9 hours ago

FBI raids of LAUSD Supt.'s home and office appear tied to AI chatbot probe

The FBI conducted a raid on the Los Angeles Unified School District (LAUSD) headquarters, executing search warrants as part of an ongoing investigation. The article examines the details surrounding the FBI's actions and the potential implications for the school district.

latimes.com
4 0
Summary
'Probably' doesn't mean the same thing to your AI as it does to you
colinprince about 12 hours ago

'Probably' doesn't mean the same thing to your AI as it does to you

The article explores how the word 'probably' is interpreted differently by AI systems compared to humans, highlighting the challenges in communicating uncertainty and ambiguity between humans and machines.

theconversation.com
4 0
Summary