1. I click the new "ToDos" tab in slack 2. the 2nd todo is from April ???????

Let’s talk money. Literally. ChatGPT Voice has come to Finances. Securely connect your accounts through Finances, then use Voice to talk spending, savings, and all the money questions on your mind. Try it: chatgpt.com/finances

Side chats are no longer ephemeral

How I give my dot responsibilities: "I'm giving you the responsibility to check for any updates we need to make to dots documentation. Go ahead and draft PRs, check once a day, do this indefinitely." specify - trusted actions - how frequently - for how long

Pro tip: just send the link to this article directly to your dot and it'll create a responsibilities page for you Then go into that Page and delete the ones you don't want, and ask dot to add the ones you do want

This was more well-received than expected -- people are definitely experiencing 2 issues with personal agents of all types: 1. overlapping/convoluted responsibilities 2. mutual lacking clarity of responsibilities

one of the most interesting things about using a personal agent like dots through imessage is that it truly stops feeling like software. as if something is entering your world as opposed to you entering its which is almost every other product. there is this visceral & maybe subtle feeling behind it that is quite powerful. a lot of ppl don’t like it but damn it’s so easy to use, & feels so damn natural esp when ai uses emojis + reactions well etc. kinda reminds me of how even babies intuitively are able to figure out how to unlock an ipad.

if you don’t feel like your agents are working for you. wipe the slate clean and use this. be open to the possibility it was a skill issue. just trust me.

People are generally getting two benefits from personal agents: 1. Small moments of delight 2. Prompted momentous efforts This is how you introduce a new category: 3. Ongoing responsibilities

chad

If there was a bear case, it would be this I've grown analog with my other productivity systems to help mitigate what is 100% true here

Sam Altman quote from the Cheeky Pint podcast Sam on their internal model - "There was one weekend when we got a new model, Im not supposed to say this, but I will, The name of the channel was [model name]-hold-my-beer... People were just throwing at it the hardest problems they could... hardest research problems from their academic career, the model was just knocking them down one after the other across many different fields... I was up most of the night for two nights in a row because I was so excited. Many of these problems we published"

Double it and give it to the next person

Today, @JudgmentLabs opens for general access. We believe long running agent systems will generate more exhaust data than the world can currently process. Hidden within agent traces are rich records of how to make these systems better: what users ask for, when agents stumble, and how agents interact with the external world. Judgment is the most performant monitoring and improvement layer for long-horizon agents. We’re excited to see what you build!

Yes

yes! finally someone else is talking about it. dot + 24x7 mac is the elite setup. you can literally call your dot and with only voice do ANYTHING on your computer. just by chatting. hyped for the blog post.

This is the future

Being able to text my dot on iMessage is really slick, but that's not why it's one of my favorite features we've released with dots so far You can tell your dot to reach you in different channels for different reasons -- chatGPT for reg updates, slack for work-related stuff But I have my dot text me (via iMessage) for anything urgent. I'm not always reachable by slack, I'm not always checking my chatGPT notifications, but if you need to reach me urgently, shoot me a text Try this after setting up iMessage or RCS on Android: "For any urgent matters, always prefer to send me a text via iMessage or RCS"

"use remotion" "use hyperframes" "use ffmpeg" Astra: "I extracted the cursor movements pixel by pixel" Just ask for outrageous things and the model will figure it out, it's that simple. These frameworks are great but are getting bitter-lessoned