news newest ask show jobs

Open Source @Github

fp.

CFTC Advisory on Self-Certification of Incentive Programs for Prediction Markets

https://www.cftc.gov/PressRoom/PressReleases/9282-26

1•petethomas•47s ago•0 comments

Living with Depression: The Part I Never Say Out Loud

https://medium.com/freedomofthought/living-with-depression-the-part-i-never-say-out-loud-3e9c218b...

1•raynchad•59s ago•0 comments

What Kind of Particularism?

https://medium.com/freedomofthought/what-kind-of-particularism-489c113a8a9c

1•raynchad•2m ago•0 comments

Lossless codec for AI agent messages – 36% fewer tokens, overhead counted

https://github.com/reh8n/a2acompress

1•reh8n•3m ago•0 comments

Who invented the great numerical algorithms? [pdf]

https://math.pku.edu.cn/teachers/litj/notes/numer_anal/inventorstalk.pdf

1•nill0•3m ago•0 comments

My First Post

1•VimalRB•5m ago•0 comments

Show HN: CoreTrace, a visual 16-bit CPU simulator

https://coretrace.srianjaneyam.me/

1•mr-anjaneyam•18m ago•0 comments

AI to Kids by Letting Them Tweak Local Chatbots [video]

https://www.youtube.com/watch?v=fNkt41q7eqs

2•dbreunig•23m ago•0 comments

Rsync 3.5 Released as "Extraordinary" Update to Fix 33 Security Issues

https://www.phoronix.com/news/Rsync-3.5

1•Bender•23m ago•0 comments

Long Story Short: Daily word game

https://gizmodo.com/long-story-short-wants-to-be-the-wordle-of-editing-bad-sentences-2000797842

1•akashwadhwani35•25m ago•0 comments

General Dynamics Failed to Build Artillery Shells for the Army

https://www.propublica.org/article/general-dynamics-artillery-factory-failed

3•colinprince•28m ago•0 comments

Is Bending Spoons Good for Startups?

https://s-1.vercel.app/posts/bending-spoons-a-capital-cycle-play/

2•Altaba•28m ago•0 comments

Show HN: UTC Time - live clock, ISO 8601, Unix timestamp

https://utctime.app/

1•nadermx•33m ago•0 comments

DeepSeek V4 Pro 0813: Intelligence, Performance and Price Analysis

https://artificialanalysis.ai/models/deepseek-v4-pro

1•theanonymousone•34m ago•0 comments

Install Manticore Search with one command

https://medium.com/@s_nikolaev/install-manticore-search-with-one-command-8ab5ba07a06b

1•snikolaev•39m ago•0 comments

Specula: Scaling formal specs for autonomous model checking of system code

http://muratbuffalo.blogspot.com/2026/08/specula-scaling-formal-specifications.html

1•matt_d•42m ago•0 comments

16GB EGPU for AI Inference, $699

https://www.amazon.com/GIGABYTE-Graphics-128-bit-Server-Grade-Thunderbolt/dp/B0GKFHSPN8

2•shrubble•42m ago•3 comments

Show HN: I asked Claude to write DDIA as a live comic

https://systemscomic.com/ddia/read

2•samuelxing•43m ago•0 comments

How the smiley face was born

https://www.history.com/articles/smiley-face-history

2•inatreecrown2•48m ago•0 comments

Edge is dropping older extensions(Manifest V2), affecting popular privacy tools

https://www.malwarebytes.com/blog/news/2026/08/edge-is-dropping-older-extensions-affecting-popula...

7•BearBest•53m ago•0 comments

You're Absolutely Right

https://www.lesswrong.com/posts/u8TdDutDyaSxG76hn/you-re-absolutely-right

1•LinchZhang•1h ago•1 comments

Show HN: Rule-lite – Tiny dependency-free TS engine for JSON conditional rules

https://github.com/tejas821/rule-lite

1•tejas821•1h ago•0 comments

From assistance to execution: How enterprises put AI to work

https://openai.com/index/how-enterprises-put-ai-to-work/

2•vcfgdev•1h ago•0 comments

A unified CLI for native OS keychains – x keyring

https://www.x-cmd.com/blog/260811/

1•Zhengqbbb•1h ago•0 comments

OpenLiving - A data center you can live in [video]

https://www.youtube.com/watch?v=BwkBXNOPa_w

1•inatreecrown2•1h ago•0 comments

Pixel Watch 5 adds a feature Apple Watch desperately needs

https://www.cultofmac.com/news/pixel-watch-5-insulin-resistance-trends

4•ValentineC•1h ago•1 comments

Examples for use of AI and especially LLMs in major mathematical developments

https://mathoverflow.net/questions/502120/examples-for-the-use-of-ai-and-especially-llms-in-major...

1•vismit2000•1h ago•0 comments

Show HN: Agent Memory Leaderboard – first public results for AI memory systems

https://agentmemoryleaderboard.ai/leaderboard/academic/textual

2•IreneAI•1h ago•0 comments

SLAC: CPU-to-GPU Side-Channel Attacks via System-Level Cache on Apple Silicon

https://arxiv.org/abs/2608.09075

3•sbulaev•1h ago•0 comments

Clamiga: Common Lisp for the Amiga

https://nnamgreb.de/blog/Clamiga+-+Common+Lisp+for+the+Amiga

3•emptybits•1h ago•0 comments

Open in hackernews

16GB EGPU for AI Inference, $699

https://www.amazon.com/GIGABYTE-Graphics-128-bit-Server-Grade-Thunderbolt/dp/B0GKFHSPN8

2•shrubble•42m ago

Comments

shrubble•42m ago

you could have 2 to get 32GB VRAM, which is enough to run a decent AI model locally.

FloatArtifact•24m ago

You're better off with a single card, like intel arc pro b70 with 32 gigabytes of VRAM unless you need something external. You're going to lose quite a bit of performance over thunderbolt. However, Nvidia still is king for its ecosystem.

shrubble•18m ago

If you could get 2x 16GB eGPU in external cases for $1398 vs the Intel B70 for $1199 to $1299 (MicroCenter pricing), some would choose the ease of use of getting the Nvidia solution up and running, over the lower price. Personally I'm rooting for Intel to be stronger in the GPU department.