frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

My falling-out with the rationalist community

https://lcamtuf.substack.com/p/my-falling-out-with-the-rationalist
1•surprisetalk•3m ago•0 comments

Show HN: Velprium Workspace

https://www.velprium.com/
1•mjkl7896•3m ago•0 comments

Mysterious little red dots at the beginning time may have an explanation

https://www.livescience.com/space/mysterious-little-red-dots-at-the-beginning-time-may-finally-ha...
1•arto•5m ago•0 comments

Ramp AI Router

https://router-website-ramp.vercel.app/router
1•robbiet480•6m ago•1 comments

Poverty of attention. Text became free. Attention didn't. (German)

https://christopher-helm.com/aufmerksamkeit-ueberfischung-ki-journalismus/
1•chelm•7m ago•0 comments

Repos in 29 days with an AI pipeline – what broke wasn't code

https://lkforge.com/blog/26-repos-in-29-days/
1•lucianBKR•7m ago•0 comments

US judge approves Anthropic's $1.5B settlement of copyright lawsuit

https://www.reuters.com/world/us-judge-approves-anthropics-15-billion-settlement-copyright-lawsui...
1•Anon84•9m ago•0 comments

Free LLM balancer combines multiple local inference machines with cloud fallback

https://github.com/Gysho/LLMrPro
1•ThomasWink1981•9m ago•1 comments

DeskPad – Virtual Keyboard and Mouse for Android Desktop Mode

https://github.com/auxiliaryutils/deskpad-app
1•Nemeses1•10m ago•1 comments

Confront Reality (2012)

http://www.aaronsw.com/weblog/anders
2•waaldev•10m ago•0 comments

An Ode to the Old Ways

https://agniva.me/personal/2026/07/19/ode-to-old-ways.html
1•sangeeth96•11m ago•0 comments

Richard Hamming: how one frustrated mathematician made data repair itself

https://nirmalutwani.substack.com/p/when-bits-flip-and-everything-breaks
1•nutwani91•11m ago•1 comments

Did California's DMV Director lie to the legislature?

https://papersplease.org/wp/2026/07/01/did-californias-dmv-director-lie-to-the-legislature/
1•Jimmc414•11m ago•0 comments

New Orleans Cops Published Policy Document Allowing Weaponized Drones

https://www.404media.co/new-orleans-cops-published-policy-document-allowing-weaponized-drones/
2•Jimmc414•12m ago•0 comments

The depth problem with agentic research

https://moekhalil.substack.com/p/the-depth-problem-with-agentic-research
1•mfkhalil•13m ago•0 comments

Trump imposes 50% tariff on Canadian goods

https://apnews.com/article/trump-canada-50-tariff-autos-alcohol-cheese-ca268ef24992b16695954c4c2e...
1•bradleyankrom•14m ago•1 comments

AI Voice Phishing Performs on Par with Human Scammers at a Fraction of the Cost

https://fredheiding.substack.com/p/ai-voice-phishing-performs-on-par
1•DalasNoin•15m ago•0 comments

Stealth Crawlers Are Not a Threat to the Open Web. Bills Targeting Them Would Be

https://www.eff.org/deeplinks/2026/07/stealth-crawlers-are-not-threat-open-web-bills-targeting-th...
2•hn_acker•15m ago•0 comments

Inference as Stereotype

https://hollisrobbinsanecdotal.substack.com/p/stereotype-threat
1•HR01•16m ago•0 comments

Does "rtk" skill cut agent tokens by 60–90%? We tested it

https://blog.jetbrains.com/ai/2026/07/rtk-claude-code-token-savings/
1•kindaAnIdiot•17m ago•0 comments

D-FINE-seg – detection, instance and semantic segmentation in one model

https://github.com/ArgoHA/D-FINE-seg
1•argo_cve•18m ago•0 comments

Cognition acquires TierZero to extend Devin into production operations

https://runtimewire.com/article/cognition-acquires-tierzero-devin-production-operations
1•ryanmerket•19m ago•0 comments

Logical Foundations of Prolog

https://www.youtube.com/playlist?list=PLEvH6T-1oh76mgfZm5QbLQoRsyr87BhiR
2•b-man•19m ago•0 comments

Aipac spends millions in Michigan Senate race – AP News

https://apnews.com/article/aipac-michigan-senate-stevens-elsayed-588bb869eb557643950bf68d6a0d7a10
4•abdelhousni•22m ago•0 comments

Show HN: Forall – OpenSpec but with actual machine-checkable verification

https://github.com/astrio-labs/forall
1•Nolan_Lwin•22m ago•0 comments

Trump to hit Canada with new 50% tariff

https://www.cbc.ca/news/world/trump-canada-us-trade-tariffs-cusma-usmca-9.7276973
3•llm_nerd•23m ago•0 comments

Ramp opens AI model router, says it cut internal LLM costs 30%

https://runtimewire.com/article/ramp-router-ai-model-gateway-llm-costs
1•ryanmerket•25m ago•0 comments

Esoterica: Prototype Game Engine

https://github.com/BobbyAnguelov/Esoterica
2•wertyk•26m ago•0 comments

The Artisanal RPG Culture of Classic Fallout Modding

https://tripharrison.substack.com/p/the-artisanal-rpg-culture-of-classic
3•droptablemain•27m ago•0 comments

Own Your Weights

https://twitter.com/jaminball/status/2075602206099919159
2•gmays•27m ago•0 comments
Open in hackernews

Apple Is the King of AI and Nobody Knows It

https://limitededitionjonathan.substack.com/p/apple-is-the-king-of-ai-and-nobody
5•jinjin2•18h ago

Comments

VaderAi•17h ago
Apple are cashflow positive, and beginning of 2026 they launched apple intelligence . they are currently working with Gemini. Apple is not the king of AI — Google, OpenAI, and Anthropic are ahead.
WheelsAtLarge•17h ago
My bet is that 99% of people will never want to deal with installing an AI model on their machine. It's not worth their time and effort. Most people will get their AI from one of the hyperscalers or a server somewhere in the cloud. Think email, Gmail is a pain sometimes but we use it simply because it's so easy to get an account and not deal with maintenance. There will be a time, soon, where a good enough AI will be good enough for most people. A whatever company will have a cheap service you can buy. That's it. It's true that you can buy a cheap Apple setup relative to Nvidia but very few people want to deal with the headache of maintaining a model.

I suspect there will be a time when we will be able to install a full model in an app in an iPhone but that future is not here yet and in that future there will be other reasons why you will still want an account on the cloud. Prices will adjust accordingly. Nvidia will likely be part of that future. It's not going anywhere.

GuestFAUniverse•11h ago
From my POV all Mac with a relevant amount of RAM (>64GB) are vaporware: I cannot order any single one. MacBooks up to 64GB are available (nice, but not even close to the machines mentioned in the article).

I can order multiple PRO 6000 from trusted sellers, immediately.