Mete Polat

New achievement unlocked:

So Dot will solve captchas and login for you with your authorization but it refuses to use a sign-in code sent to the email it has access to. What?

The only thing standing between me and AGI-pilled life is Claude signing me out randomly every couple of weeks.

I’ve thought about this a lot myself (cause I’m in on the mania). I don’t have an answer. But I think it’s interesting to consider where’s the line between unproductive avoidance vs creative expression. How many musicians have picked up their instruments to escape emotional pain only to channel it through whatever they create? This applies to basically any creative. I think it’s important to be self-aware enough to not completely ignore root causes of your discomfort but sometimes creating things is the best way to come to terms with and channel whatever you’re feeling. And yes, that involves vibe coding.

Actually interesting to see this cause on the other end it keeps messaging me multiple times a day with stuff I don’t care about or already took care of (unlike Instinct or Dot) so they’re just loading up the engagement mechanics on the proactive end vs reactive end

I’m going in a 5-week Asia trip in November and doing it one-bag style. Strongly considered leaving my laptop behind but I need it for occasional work and writing. With how good Codex computer use has gotten, I can honestly see myself just leaving it always-on at home and using Dot to do everything remotely. The only thing stopping me is that it still routinely disconnects or Apple throws some permission prompt you have no way to resolve without finicky screen-sharing. But it’s almost at the point where I could do that which is amazing. Apple is honestly the biggest bottleneck here. That’s why we need OpenAI hardware.

This. NYC adds one key ingredient to your life which you’ll be hard-pressed to find anywhere else - serendipity. To me it’s a must-have ingredient for a happy life. And it’s basically absent from any non-urban area or anything too spread out (and hence requiring a car). NYC is a top-tier serendipity generator.

Something that @ssankar said in his @stratechery interview (highly recommend it) resonated with me because I think the same can be said about the design industry: “The bigger issue is the profound lie of globalization. The calculus that globalization got wrong is this idea that we can do the innovation and “they” can do the production — “they” make the things, we do the innovation. The problem is that innovation is a consequence of productivity, if you stop making the thing, you cede the stimulus required to do the innovation around the thing.” In the last decade+, designers & design orgs broadly increasingly distanced themselves from engineering and business sides while maintaining the illusion of being the main drivers of innovation for the business. This where I see AI improving things greatly as the line between design, engineering, and product is increasingly getting blurred.

Been using Parcel app for years to track packages but often just forgot / was too lazy to add stuff. Now I hooked it up to my Dot and whenever it detects a shipment notification email, it adds it to my app and then I get all tracking notifications. Pretty neat. apps.apple.com/app/id3755892…

The new MacBook Pros with a touchscreen are just another ploy to sell us more polishing cloths.

My conspiracy theory: Rockstar Games actually invented LLMs to wipe out white collar labor and created a giant pool of disenfranchised unemployed men with infinite free time - just in time for GTA VI release.

You can now set your Dot as the ChatGPT launch screen. Great to see the team addressing feedback quickly. It still doesn’t work from the iPhone Action button for some reason. Hope that’s fixed soon. It’s basically my primary launch path.

How can I short Instinct @ 30B?

Apple 🤝🏼 Me both featuring The Strokes's new album in our promos.

I've been using AI as my fitness (calisthenics) coach for the past 6 months. My workflow evolved quite a bit. I started with a dedicated Hermes agent that used Obsidian as its context base. We created a program together and then it would create workouts in a note whenever I asked. Both Hermes and Obsidian were too clunky to keep this up but I was hooked on the general value of hyper-personalized fitness programming Now, I have my Dot writing my workouts into Craft notes, using the context from my program and a movement library we created (with nice images it generated for each of them). Each day I tell Dot if I'm working out today and it links my workout in my daily note. It works super well. The main challenges I keep running into: 1. Providing enough feedback to my agents so they can keep adapting the program - I now log RPE for each set and leave free text notes after a workout (injury signals, programming preferences, etc) 2. A better presentation & logging interface. Craft is almost perfect for my needs but I'm experimenting with my own app that uses a 3D model to show movements & better logging UI. I expect agents will get good enough to generate this UI perfectly on the fly within a year. Overall, it's been a game-changer. AI can actually understand account for all the intricacies like injuries, schedules, travel, equipment availability, goals, programming preferences, etc. I was constantly thrown off by injuries & struggled to return to my routine. AI helped me figure out movement patterns that will help me achieve my goals while preventing future injury. Obviously, I have quite a bit of knowledge of what I need & want through prior experience & PT, so I can steer. More importantly, I can do this on my terms without having to descend into full datamaxxing mode of tracking every single metric, readiness etc. I only track key metrics that move me towards my goals, the rest is vibes. I'm on track for my goals this year of 10s free handstand, 20s L-sit, and 1 clean pistol squat (2 year journey mired by back injuries so setting the goal low).

If you're using Dots to spin up Codex threads, it doesn't know which model or effort its using and has little to no control over it. It worked a few times if you specify it upfront but for the most part it seems completely aloof.

Many corporations and companies chose to abstract themselves from their customers through terrible interfaces, unreachable or non-existent customer support, and dark patterns designed to confuse you. This didn’t happen out malice. These are forces of scale, profit incentives, and constraints. But for every action, there is an equal and opposite reaction. Agents are a counter-force. Consumers can now deputize a persistent entity advocating & acting on their behalf. These forces of abstraction are now available to an average consumer.