Don’t mistake kindness for weakness

Coming soon

@lessin and everyone pods

now everyone types... and there are no more professional typists - soon everyone codes, and there will be no professional developers.

@timyangnet @johnpalmer this is a valid point and one that rattles around in my head as well... the moat is brooked easily.

@johnpalmer I’m looking at this from a different angle than the author, though. As an AI user, how much of a lead do early adopters actually have over latecomers? I feel like any "understanding advantage" they have could be closed in a single week.

@atmoio Don’t remember him mentioning “sentience” but will reread

The real virus is articles like this.

@joelcomm did you use ai to help write this?

@mattshumer_ @HellenicVibes how do you know they are your original, brain-prompted thoughts? serious question.

@HellenicVibes These are 100% my thoughts... AI helped organize them and improved my writing. You don't need to outsource everything to AI to get value from it :)

@AlexFinn What are you working on?

getting a lot of q's around codex & opus the new codex is great. the desktop app is excellent. I use it daily Idk what it is, Opus just feels so eager and excited to build with me. hard to put my finger on it. you honestly can't lose with either tho codex great value for $$

Building with Claude Opus 4.6 in Claude Code is literally a holy experience It feels like I have unlimited power. Anything I think just appears in front of me

@AlexFinn Cool. What has been achieved by the company thus far?

Saturday night. 6 hours of sleep over the last week. My autonomous agent company having an emergency meeting on the left. My ClawdBot giving them new tasks on the right All being powered by local models in my Mac Studio data center I refuse to be in the permanent underclass

@TheHoolest This is practically a dark pattern. It's at the very least rather annoying ... If the consumer simply wants the hardware and not the subscription, it's apparently not possible despite what the checkout flow initially indicates?

@minchoi Is this what Proto does?

If you enjoyed this thread, Follow me @minchoi and please Bookmark, Like, Comment & Repost the first Post below to share with your friends: x.com/minchoi/status…

It was only Day 1 of CES 2026. And the announcements are already mind blowing. 10 wild reveals: 1. Samsung's mysterious slim 3D display

@Dulles_Airport Just focus on getting flights in and out on time. That’s an area of improvement for you in 2026. Add that to your goals

Good morning to everyone going back to work today and trying to remember their logins! Stay strong. You got this!

@hereisalden #aldn fans, check this shirt I got for Christmas- look real close

@Wisdom_HQ Wisdom is unhappy

@Wisdom_HQ 2

@AIExplainedYT @karpathy do you have the models consulting on the best outcome?

@karpathy Hey Andrej, what do you think of my version?: lmcouncil.ai More like 6 months than one weekend, as I expanded the concept into images, audio, polls, smartgpt leaders and more.

As a fun Saturday vibe code project and following up on this tweet earlier, I hacked up an **llm-council** web app. It looks exactly like ChatGPT except each user query is 1) dispatched to multiple models on your council using OpenRouter, e.g. currently: "openai/gpt-5.1", "google/gemini-3-pro-preview", "anthropic/claude-sonnet-4.5", "x-ai/grok-4", Then 2) all models get to see each other's (anonymized) responses and they review and rank them, and then 3) a "Chairman LLM" gets all of that as context and produces the final response. It's interesting to see the results from multiple models side by side on the same query, and even more amusingly, to read through their evaluation and ranking of each other's responses. Quite often, the models are surprisingly willing to select another LLM's response as superior to their own, making this an interesting model evaluation strategy more generally. For example, reading book chapters together with my LLM Council today, the models consistently praise GPT 5.1 as the best and most insightful model, and consistently select Claude as the worst model, with the other models floating in between. But I'm not 100% convinced this aligns with my own qualitative assessment. For example, qualitatively I find GPT 5.1 a little too wordy and sprawled and Gemini 3 a bit more condensed and processed. Claude is too terse in this domain. That said, there's probably a whole design space of the data flow of your LLM council. The construction of LLM ensembles seems under-explored. I pushed the vibe coded app to github.com/karpathy/llm-c… if others would like to play. ty nano banana pro for fun header image for the repo

@tbpn @johncoogan @jordihays

@MoreorLessPod @davemorin @brit @lessin @Jessicalessin people seemed to like the idea when I posted (and tried to build) a couple years ago. It’s Apple’s game to lose, really… bots in iMessage and group chats? Killer app

@HumaneBlanks got notified something i'd long been waiting for was back in stock and the text link goes here. Please fix humaneblanks.com/products/900-g…

@tbpn @jordihays @johncoogan Gemini in iMessage, not the other way around messagepals.com

@minchoi When Apple wakes up and enables this in iMessage they will finally be able to player

NEW: group chats in ChatGPT just dropped. Drop friends + AI in one convo to plan trips, settle debates, and auto-summarize chaos. Sam wants ChatGPT to be the everything app🤔