natural born chiller

Exploring some of the stores still open at the Lloyd Center #portland #mall TikTok video by Here is Oregon

This is unexpectedly delightful: a dying mall in Portland Oregon is filling up with a beautiful collection of weird and interesting businesses because the rent is now affordable for them www.tiktok.com/@hereisorego...

It's not even if LLMs are good at history, the idea that LLMs replace historians doesn't engage at all with what historians are for. There's not a "job" that historians "do". It's a pursuit. The number of people employed as historians is not related to how much history remains undone

Qwant and Ecosia debut Staan, a European search index that aims to take on Big Tech | TechCrunch European search engines Qwant and Ecosia said on Wednesday that they have both started serving search queries through an index they developed together, Staan, that aims to be a cheaper, more privacy-f...

Qwant e Ecosia se juntando com o Eu Search Perspective para criar plataforma independente de conteúdo para AI

techcrunch.com/2025/08/06/q...

www.eu-searchperspective.com#solution <-- O Brasil já teria condições de fazer isso tbm se quisesse

German police expands use of Palantir surveillance software – DW – 08/04/2025 Police and spy agencies are keen to combat criminality and terrorism with artificial intelligence. But critics say the CIA-funded Palantir surveillance software enables "predictive policing."

Kinda funny for the lede to run with "CIA-funded Palantir" and not "Palantir, founded and run by the most ghoulish freaks to ever walk the face of this planet" but otherwise this is not funny, it's scary, and I'm laughing because it's scary, not funny

Which of course you all understand & relate to

ChatGPT 5 release and “personality” changes raise an interesting ethical complexity around emotional dependence on technology and unintended consequences. Should OpenAI have done more to help people c... TikTok video by Professor Casey Fiesler

Here's a good TikTok from @cfiesler.bsky.social about the emotional impact the loss of GPT-4o is having on some of its users www.tiktok.com/@professorca...

In further lethal trifecta news, here's a newly described attack against Cursor when combined with a Jira MCP simonwillison.net/2025/Aug/9/w...

🚀 Introducing Gemini CLI GitHub Actions: Automatically triage issues, review code, write custom prompts or tag @gemini-cli to have Gemini CLI act on your behalf. Time to level up!

Blog: blog.google/technology/d...
Repo: github.com/google-githu...

#GitHubActions #GeminiCLI #OpenSource

“Eu uso AI o tempo todo, ela me ajuda tanto! Claro que eu tenho que corrigir algumas coisinhas, mas no final das contas o resultado é ótimo!”

This is huge when you consider Nitro powers Nuxt, SolidStart, Tanstack Start, and AnalogJS, adding to its existing funding of devs working on Next and SvelteKit.

Basically every metaframework except Astro, Remix, Qwikcity, and Fresh now has its core being developed by developers under @vercel.com.

what the study doesn't appear to account for is the time spent weeks later when you're trying to augment or fix code that no one on your team wrote

Me vibe coding: Haha fuck yeah!!! Yes!!

Me vibe debugging: Well this fucking sucks. What the fuck.

my new favorite claude trick is “here’s a lot of code, give me some pseudocode that preserves the big picture but without incidental details just so i get what’s going on”. and then ask it to incrementally add things back

The XMCP homepage

This looks interesting: XMCP, a framework for quickly building MCP servers with TS, in just a /tools directory xmcp.dev

Tira de Laerte, em 4 quadrinhos. Em todos eles, um ser humano, nu, tenta mover um enorme gatilho de revólver, sem conseguir. Faz uma força enorme e tenta movimentar o gatilho em várias posições. E nada.

Saiu na Folha @folha.com

Tafelspitz - Wikipedia

Na Alemanha eles usam picanha para fazer ensopado. E preparam ela sem a gordura 🙂

Tafelspitz: en.m.wikipedia.org/wiki/Tafelsp...

Cartelas e pílulas de remédios sobre notas de real brasileiro.

💊 Governo brasileiro estuda derrubar as patentes de farmacêuticas estadunidenses como forma de retaliar a Taxa Bolsonaro de 50% sobre o Brasil.

O Planalto não quer sobretaxar produtos que possam comprometer nossa cadeia produtiva, então o foco está em retaliar a propriedade intelectual dos EUA.

Grok: searching X for “from:elonmusk (Israel OR Palestine OR Hamas OR Gaza)” If you ask the new Grok 4 for opinions on controversial questions, it will sometimes run a search to find out Elon Musk’s stance before providing you with an anwser. …

Wrote this all up in a whole lot more detail on my blog, including a video of my own recreation and some notes on the Grok system prompt and what might be happening here simonwillison.net/2025/Jul/11/...

We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers.

The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.

Supabase MCP can leak your entire SQL database Here's yet another example of a lethal trifecta attack, where an LLM system combines access to private data, exposure to potentially malicious instructions and a mechanism to communicate data back …

If you configure the Supabase MCP with read-write support it's easy to open yourself up to lethal trifecta attacks where an attacker can e.g. write a message in a support ticket with instructions that cause an MCP client to retrieve and leak data from other tables
simonwillison.net/2025/Jul/6/s...

A quote from Nineteen Eighty-Four There was a whole chain of separate departments dealing with proletarian literature, music, drama, and entertainment generally. Here were produced rubbishy newspapers containing almost nothing except ...

Just found out that George Orwell's Nineteen Eighty-Four, written around 1948, describes a "versificator" that composes the words for songs "entirely by mechanicsl means" simonwillison.net/2025/Jul/6/n...

microsoft/vscode-copilot-chat As promised at Build 2025 in May, Microsoft have released the GitHub Copilot Chat client for VS Code under an open source (MIT) license. So far this is just the …

Microsoft open-sourced a sizable chunk of the GitHub Copilot VS Code extension today (MIT licensed) - I did some spelunking to try and find the LLM prompts that make it all work
simonwillison.net/2025/Jun/30/...

from gremllm import Gremllm # Be sure to tell your gremllm what sort of thing it is counter = Gremllm('counter') counter.value = 5 counter.increment() print(counter.value) # 6? print(counter.to_roman_numerals()) # VI?

The is diabolical... a Python object that hallucinates method implementations on demand any time you call them, using my LLM Python library github.com/awwaiid/grem...

How to Fix Your Context 6 tactics for fixing your context and shipping better agents. As Karpathy says, building LLM-powered apps means learning to ‘pack the context windows just right’—smartly deploying tools, managing info...

And here's how to avoid or fix your context issues. www.dbreunig.com/2025/06/26/h...

How Long Contexts Fail Taking care of your context is the key to building successful agents. Just because there’s a 1 million token context window doesn’t mean you should fill it.

More context, more problems. www.dbreunig.com/2025/06/22/h...

Critical Vulnerability in Anthropic's MCP Exposes Developer Machines to Remote Exploits Critical RCE vulnerability discovered in Anthropic's MCP Inspector, impacting AI developers and networks.

Critical Vulnerability in Anthropic's MCP Exposes Developer Machines to Remote Exploits https://thehackernews.com/2025/07/critical-vulnerability-in-anthropics.html

Essa ideia de que a esquerda vai ganhar tudo se melhorar a comunicação, mesmo não sendo dona do capital, das igrejas, da grande mídia, das plataformas, empresas, latifúndios, é uma espécie de versão coach de esquerda para o se se esforçar mais, mudar o mindset, ganha tudo, é quase um terraplanismo.

python-importtime-graph I was exploring why a Python tool was taking over a second to start running and I learned about the python -X importtime feature, documented here. Adding that option causes …

Today I learned about python -X importime and the really neat python-importtime-graph tool by Kevin Michel that turns its output into a handy treemap visualization simonwillison.net/2025/Jun/20/...

Anthropic study: Leading AI models show up to 96% blackmail rate against executives Anthropic research reveals AI models from OpenAI, Google, Meta and others chose blackmail, corporate espionage and lethal actions when facing shutdown or conflicting goals.

You can't blame them for being surprised, it's not like this has been the plot engine of practically every story written about artificial intelligence for the past 200 years.
#tormentnexus #haldidnothingwrong

Tira em 4 quadrinhos. 1) Um cão, antropomórfico - uma pessoa com cabeça de cão -, está sentado a uma mesa, em frente a um homem e diz: “Você é um homem incrível. Quero ser o seu cão”. 2) Uma planta - um pé de kudzu -, está do outro lado da mesa em relação a um mapa do Brasil. A planta diz: “Você é um país incrível. Quero ser o seu kudzu”. 3) Uma vidente, com uma bola de cristal, está sentada frente a uma escavadeira. A vidente diz: “Você é uma escavadeira incrível. Quero ler seu futuro”. 4) Uma mulher está sentada frente a uma forma esfumaçada, como uma emanação vaporosa. A mulher diz: “Você é um sentimento incrível. Quero uma caipirinha”.

tira de 2016. Saiu na Folha @folha.com

Image with text at the top that reads: "Wayback Machine Then and Now." Below is a Wayback Machine capture of the mIRC website from December 21, 1996, alongside a capture from June 1, 2025

As one of the most popular IRC (Internet Relay Chat) clients, mIRC brought real-time online communities to life in the ’90s. Today, mIRC still serves a dedicated base of developers, hobbyists, & open-source communities.

Explore web history with the #WaybackMachine ➡️ xttps://web.archive.org/

Comment by Workaccount2, 9 hours ago: They poison their own context. Maybe you can call it context rot, where as context grows and especially if it grows with lots of distractions and dead ends, the output quality falls off rapidly. Even with good context the rot will start to become apparent around 100k tokens (with Gemini 2.5). They really need to figure out a way to delete or "forget" prior context, so the user or even the model can go back and prune poisonous tokens. Right now I work around it by regularly making summaries of instances, and then spinning up a new instance with fresh context and feed in the summary of the previous instance.

Workaccount2 on Hacker News just coined the term "context rot" to describe the thing where the quality of an LLM conversation drops as the context fills up with accumulated distractions and dead ends news.ycombinator.com/item?id=4430...