Mete Polat
@metedata.bsky.social
Product Design @ Netflix, previously Peloton. Immigrant, tinkerer, prolific walker, learner, designer, ramen lover.
The masculine urge to create a to-do app
www.theverge.com/2025/1/14/24...
www.theverge.com/2025/1/14/24...
ChatGPT can now handle reminders and to-dos
ChatGPT just got a new automation feature for future tasks.
www.theverge.com
January 15, 2025 at 1:45 AM
The masculine urge to create a to-do app
www.theverge.com/2025/1/14/24...
www.theverge.com/2025/1/14/24...
One of the true simple pleasures that tickle something deep in our brains is a sharp knife smoothly cutting into a tomato.
January 15, 2025 at 1:44 AM
One of the true simple pleasures that tickle something deep in our brains is a sharp knife smoothly cutting into a tomato.
The most mindblowing thought experiment from @acquiredfm.bsky.social on Meta:
It would take Meta ~11 years to break even on Reality Labs investment if they were to create the most profitable and widely adopted product in human history (modeled after the iPhone) and release it tomorrow.
It would take Meta ~11 years to break even on Reality Labs investment if they were to create the most profitable and widely adopted product in human history (modeled after the iPhone) and release it tomorrow.
January 15, 2025 at 1:44 AM
The most mindblowing thought experiment from @acquiredfm.bsky.social on Meta:
It would take Meta ~11 years to break even on Reality Labs investment if they were to create the most profitable and widely adopted product in human history (modeled after the iPhone) and release it tomorrow.
It would take Meta ~11 years to break even on Reality Labs investment if they were to create the most profitable and widely adopted product in human history (modeled after the iPhone) and release it tomorrow.
If OpenAI makes o3 available only on their ChatGPT Pro subscription (not unlikely given the high compute costs they’re hinting at), they’ll make paying $200/mo for a subscription mainstream.
Similar to how Apple normalized paying $1000+ for a phone with iPhone X.
Similar to how Apple normalized paying $1000+ for a phone with iPhone X.
December 21, 2024 at 5:54 AM
If OpenAI makes o3 available only on their ChatGPT Pro subscription (not unlikely given the high compute costs they’re hinting at), they’ll make paying $200/mo for a subscription mainstream.
Similar to how Apple normalized paying $1000+ for a phone with iPhone X.
Similar to how Apple normalized paying $1000+ for a phone with iPhone X.
Innovation is combinatorial by nature.
So the real magic of all these new crazy capabilities (like the new video models) is not what they can do on their own.
It’s how they’ll stack together with millions of other use cases, tools, processes, and ideas to create entirely new possibilities.
So the real magic of all these new crazy capabilities (like the new video models) is not what they can do on their own.
It’s how they’ll stack together with millions of other use cases, tools, processes, and ideas to create entirely new possibilities.
December 21, 2024 at 5:54 AM
Innovation is combinatorial by nature.
So the real magic of all these new crazy capabilities (like the new video models) is not what they can do on their own.
It’s how they’ll stack together with millions of other use cases, tools, processes, and ideas to create entirely new possibilities.
So the real magic of all these new crazy capabilities (like the new video models) is not what they can do on their own.
It’s how they’ll stack together with millions of other use cases, tools, processes, and ideas to create entirely new possibilities.
AI is a game-changer for language learning.
It's basically the smartest and most patient tutor on demand for any language.
It's basically the smartest and most patient tutor on demand for any language.
December 21, 2024 at 5:53 AM
AI is a game-changer for language learning.
It's basically the smartest and most patient tutor on demand for any language.
It's basically the smartest and most patient tutor on demand for any language.
A brain that reaches peak creativity at 2am is an absolute curse.
December 21, 2024 at 5:52 AM
A brain that reaches peak creativity at 2am is an absolute curse.
There has to be an app out there where you can feed your prompt to every available LLM model, then have each model compare and rate all of the responses against each other, then output a ranking based on the composite ratings.
Setting money on fire in terms of API costs, but it sounds fun.
Setting money on fire in terms of API costs, but it sounds fun.
December 20, 2024 at 6:22 AM
There has to be an app out there where you can feed your prompt to every available LLM model, then have each model compare and rate all of the responses against each other, then output a ranking based on the composite ratings.
Setting money on fire in terms of API costs, but it sounds fun.
Setting money on fire in terms of API costs, but it sounds fun.
Honestly, today's announcement from OpenAI blew my mind.
I already sort of knew about their "Work with Apps" mode, but now use their advanced voice mode in conjunction with it.
But...
www.youtube.com/watch?v=g_qx...
I already sort of knew about their "Work with Apps" mode, but now use their advanced voice mode in conjunction with it.
But...
www.youtube.com/watch?v=g_qx...
Work with Apps—12 Days of OpenAI: Day 11
YouTube video by OpenAI
www.youtube.com
December 20, 2024 at 5:38 AM
Honestly, today's announcement from OpenAI blew my mind.
I already sort of knew about their "Work with Apps" mode, but now use their advanced voice mode in conjunction with it.
But...
www.youtube.com/watch?v=g_qx...
I already sort of knew about their "Work with Apps" mode, but now use their advanced voice mode in conjunction with it.
But...
www.youtube.com/watch?v=g_qx...
If you're trying to build habits in the new year and want to maximize your chances of sticking with them, the trick is to start now to build momentum.
You don't start your hot lap from a standing position - you accelerate to the max before the start line.
You don't start your hot lap from a standing position - you accelerate to the max before the start line.
December 18, 2024 at 3:57 PM
If you're trying to build habits in the new year and want to maximize your chances of sticking with them, the trick is to start now to build momentum.
You don't start your hot lap from a standing position - you accelerate to the max before the start line.
You don't start your hot lap from a standing position - you accelerate to the max before the start line.
Throwback to the mini DIY project to display some of my favorite iPods.
December 18, 2024 at 3:57 PM
Throwback to the mini DIY project to display some of my favorite iPods.
So apparently Claude knows OpenAI's API better than ChatGPT?
I'm trying to get ChatGPT to write me a script that involves making calls to OpenAI's own API... but it keeps using outdated syntax and deprecated functions. Claude corrected to the updated syntax right away.
Fun.
I'm trying to get ChatGPT to write me a script that involves making calls to OpenAI's own API... but it keeps using outdated syntax and deprecated functions. Claude corrected to the updated syntax right away.
Fun.
December 18, 2024 at 3:56 PM
So apparently Claude knows OpenAI's API better than ChatGPT?
I'm trying to get ChatGPT to write me a script that involves making calls to OpenAI's own API... but it keeps using outdated syntax and deprecated functions. Claude corrected to the updated syntax right away.
Fun.
I'm trying to get ChatGPT to write me a script that involves making calls to OpenAI's own API... but it keeps using outdated syntax and deprecated functions. Claude corrected to the updated syntax right away.
Fun.
People ask me why I hate the MTA.
I’m currently on an A train that stopped at a Q stop but is running on an F line, while announcing on its screens that every stop is the last.
I’m currently on an A train that stopped at a Q stop but is running on an F line, while announcing on its screens that every stop is the last.
December 15, 2024 at 8:59 PM
People ask me why I hate the MTA.
I’m currently on an A train that stopped at a Q stop but is running on an F line, while announcing on its screens that every stop is the last.
I’m currently on an A train that stopped at a Q stop but is running on an F line, while announcing on its screens that every stop is the last.
Don’t know who needs to hear it but it’s ok to remove a liked song in Spotify if you no longer like it.
December 15, 2024 at 8:58 PM
Don’t know who needs to hear it but it’s ok to remove a liked song in Spotify if you no longer like it.
We should approach the inner voice in our heads like it’s our own Formula 1 race engineer.
December 15, 2024 at 8:58 PM
We should approach the inner voice in our heads like it’s our own Formula 1 race engineer.
For anyone learning Turkish, I created an Anki vocabulary deck based on “The Delights Of Learning Turkish” self-study book.
ankiweb.net/shared/info/...
ankiweb.net/shared/info/...
ankiweb.net
December 13, 2024 at 2:34 AM
For anyone learning Turkish, I created an Anki vocabulary deck based on “The Delights Of Learning Turkish” self-study book.
ankiweb.net/shared/info/...
ankiweb.net/shared/info/...
Most AI-generated vids will look off to us because we all grew up watching non-AI content. We have entrenched references in our minds.
The generation that grows up consuming AI-generated content on the regular from early childhood, the AI artifacts might feel completely normal.
The generation that grows up consuming AI-generated content on the regular from early childhood, the AI artifacts might feel completely normal.
December 13, 2024 at 2:33 AM
Most AI-generated vids will look off to us because we all grew up watching non-AI content. We have entrenched references in our minds.
The generation that grows up consuming AI-generated content on the regular from early childhood, the AI artifacts might feel completely normal.
The generation that grows up consuming AI-generated content on the regular from early childhood, the AI artifacts might feel completely normal.
Duolingo has always been optimizing for pure engagement over education.
It’s useful as a beginner to build the habit, but once you internalize the intention and discipline to learn, you’re better served by other tools.
It’s useful as a beginner to build the habit, but once you internalize the intention and discipline to learn, you’re better served by other tools.
December 13, 2024 at 2:32 AM
Duolingo has always been optimizing for pure engagement over education.
It’s useful as a beginner to build the habit, but once you internalize the intention and discipline to learn, you’re better served by other tools.
It’s useful as a beginner to build the habit, but once you internalize the intention and discipline to learn, you’re better served by other tools.
My hot take is that Apple's native Notes app sucks.
Yes, at this point it's feature-rich. Yes, it's well-integrated with iOS. Yes, it simple. Yes, it gets the job done.
But man it's so ugly and clunky to use. There's zero polish or thought given to the actual experience of writing & note-taking.
Yes, at this point it's feature-rich. Yes, it's well-integrated with iOS. Yes, it simple. Yes, it gets the job done.
But man it's so ugly and clunky to use. There's zero polish or thought given to the actual experience of writing & note-taking.
December 13, 2024 at 2:31 AM
My hot take is that Apple's native Notes app sucks.
Yes, at this point it's feature-rich. Yes, it's well-integrated with iOS. Yes, it simple. Yes, it gets the job done.
But man it's so ugly and clunky to use. There's zero polish or thought given to the actual experience of writing & note-taking.
Yes, at this point it's feature-rich. Yes, it's well-integrated with iOS. Yes, it simple. Yes, it gets the job done.
But man it's so ugly and clunky to use. There's zero polish or thought given to the actual experience of writing & note-taking.
Celebrating the murder of the UnitedHealthcare CEO is a case of confusing the symptom for the cause.
Greedy insurance corporations and CEOs that excel at extracting profit are symptoms of privatized healthcare, not the cause of it.
Greedy insurance corporations and CEOs that excel at extracting profit are symptoms of privatized healthcare, not the cause of it.
December 8, 2024 at 3:55 AM
Celebrating the murder of the UnitedHealthcare CEO is a case of confusing the symptom for the cause.
Greedy insurance corporations and CEOs that excel at extracting profit are symptoms of privatized healthcare, not the cause of it.
Greedy insurance corporations and CEOs that excel at extracting profit are symptoms of privatized healthcare, not the cause of it.
I don’t know if I hate or love that Spotify now has comments for podcast episodes.
December 1, 2024 at 12:43 AM
I don’t know if I hate or love that Spotify now has comments for podcast episodes.
My habit tracking journey coming full-circle.
I went from using one daily for years to completely abandoning it, because I noticed myself doing things just to keep the streak. I felt like it hijacked my incentives.
I’m back to it because I realized that in most cases completion > perfection.
I went from using one daily for years to completely abandoning it, because I noticed myself doing things just to keep the streak. I felt like it hijacked my incentives.
I’m back to it because I realized that in most cases completion > perfection.
November 29, 2024 at 8:43 PM
My habit tracking journey coming full-circle.
I went from using one daily for years to completely abandoning it, because I noticed myself doing things just to keep the streak. I felt like it hijacked my incentives.
I’m back to it because I realized that in most cases completion > perfection.
I went from using one daily for years to completely abandoning it, because I noticed myself doing things just to keep the streak. I felt like it hijacked my incentives.
I’m back to it because I realized that in most cases completion > perfection.
I don’t think it’s fair to compare it to OpenAI’s voice mode. There’s a big difference between a natural-sounding conversation and a conversation that can tease out the right intent and act on it. The success of the latter is binary. Neither Apple / Amazon / Google have figured it out yet.
NEW: Apple is racing to develop a more conversational version of its Siri digital assistant, aiming to catch up with OpenAI’s ChatGPT and other voice services. The company plans to introduce the revamped LLM Siri next year and launch it by spring 2026. www.bloomberg.com/news/article...
Apple Readies More Conversational Siri in Bid to Catch Up in AI
Apple Inc. is racing to develop a more conversational version of its Siri digital assistant, aiming to catch up with OpenAI’s ChatGPT and other voice services, according to people with knowledge of th...
www.bloomberg.com
November 22, 2024 at 6:07 AM
I don’t think it’s fair to compare it to OpenAI’s voice mode. There’s a big difference between a natural-sounding conversation and a conversation that can tease out the right intent and act on it. The success of the latter is binary. Neither Apple / Amazon / Google have figured it out yet.
This makes me wonder about the fate of companies like The Browser Company (makers of Arc) and others trying to build a mass-market AI-native browsing experience.
The race for a new browsing paradigm is on. And Google is not well-positioned to compete.
The race for a new browsing paradigm is on. And Google is not well-positioned to compete.
Sources: OpenAI considered making a browser, discussed deals to power AI features on Samsung devices and search on sites and apps from Condé Nast and others (The Information)
Main Link | Techmeme Permalink
Main Link | Techmeme Permalink
November 22, 2024 at 5:42 AM
This makes me wonder about the fate of companies like The Browser Company (makers of Arc) and others trying to build a mass-market AI-native browsing experience.
The race for a new browsing paradigm is on. And Google is not well-positioned to compete.
The race for a new browsing paradigm is on. And Google is not well-positioned to compete.
Everyone is raving about Bluesky on X and Threads, but after lots of “Show less like this” and careful follows, my feed is still mostly random anime, political clickbait, occasional erotica, and some recycled old tweets.
November 22, 2024 at 5:07 AM
Everyone is raving about Bluesky on X and Threads, but after lots of “Show less like this” and careful follows, my feed is still mostly random anime, political clickbait, occasional erotica, and some recycled old tweets.