Seems like an own goal, I guess fear mongering is not the best marketing strategy after all

I cannot amplify this enough.

Looks like we are going to find out if Fable’s expensive and anticompetitive launch turn out to be an own goal

I am not convinced that the solution to AI's producitivy and cost problems is a model that is atleast 2x as expensive to use as previously most expensive model to use is. It's great that it has such good performace over long-term tasks, but each of those tasks could cost.. $50 dollars in tokens, which is a lot if the task is to "make a slide deck or fix a bug" and then it still needs to be steered, because, I, the human in the loop here, do not have the ability to perfectly communicate the totality of my intent and vision is a text box.

Good point….

|￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣| | Don't Push To Production On Friday | |＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿| \ (•◡•) / \ / —— | | |_ |_

The Least Privileged Principal does not cleanly map to AI Native organizations

There is an entire industry of software that exists because Windows isn’t secure. What happens when the agents are just using Windows?

Agents, not Super Apps, are the future of AI. OpenAI and Anthropic will not be able to command their massive valuations as just model providers.

Are we doing Embrace, Extend, Extinguish again?

La magnífica humanidad que Dios ha creado se encuentra hoy ante una elección decisiva: levantar una nueva torre de Babel o edificar la ciudad donde Dios y la humanidad habiten juntos. En Jesucristo, esta magnífica humanidad encuentra el camino, la verdad y la vida, abriendo a cada uno de nosotros la vía para crecer hacia la plenitud. #MagnificaHumanitas vatican.va/content/leo-xiv/e…

Proximity to SF doesn't matter if you use Twitter Correctly. X.com is 2 months ahead of SV/founders/eng (if you curate your feed correctly and are in the right group chats)

This.

Can someone make an AI native alternative to environment variables. I spend way too much of my day copying and pasting secure variables into things I build now

Did Codex just try to HAL me?

Has anyone ever actually learned something useful from a panel discussion? Or are we all just sitting there waiting for the part where we get to talk to other interesting people and drink beer?

Colin Chapin was the OG of AI Agent Design: "Simplify, then add lightness" classicdriver.com/en/article…”

I think people like big plugins and complex agents because it makes the AI look like it's doing a lot. It runs for a long time on it's own and uses a lot of tokens. I'm not convinced it's more productive though. Pi is much faster than Claude Code, which is much faster than OpenClaw, and I think I'm more productive with an agent that returns quickly with the minimal amount of work done to achieve the goal vs a Lobster in a box that spins for 10 hours and builds a 100k line SaaS app I didn't ask for.

Ever since I was a little boy, I knew I wanted to log into Microsoft Authenticator 47 times a day.

I am very tired of cords. Even with a cordless mouse and keyboard, my desk ends up covered in cords! Each monitor needs two cords, power and display, my thunderbolt hub has a big, chunky power supply cord, and then the thunderbold cord to my macbook. Speakers, 3 cords, headphone have a little control hub which has has a cord too. I thought with Thunderbolt we were supposed to fix all this, but now I just have Thunderbolt / USB-C cords everywhere!