Anthropic's version of @thsottiaux has left. Who does Anthropic have to listen to the people of x now?!?
I’ve left Anthropic after more than a year of working on Claude Code and seeing it scale larger than I ever imagined.
I have really loved working at Ant and believe deeply in the mission, and most of all, the idea of a race to the top.
The idea that if we work hard and responsibly, we can raise the bar and inspire others to great outcomes. And if we all push each other to be better, things can be radically good. I will carry the spirit of this with me forever.
My colleagues at Anthropic are the most talented, thoughtful, competent, and earnest people I have ever met. I am glad they are at the wheel and trust them deeply.
To the community on X, thank you for everything. You’ve made this the most fun (and sometimes stressful) job I’ve ever had.
For me, it is time for a bit of a break before diving into something new.
If you want more excellent Claude Code content, you should follow @addyosmani , @lydiahallie, @_catwu, @ClaudeDevs and of course @bcherny who started it all.
0
0
0
48
Not everyone is starting from the same place with AI.
I sat down with @engstrometric to talk about asking better questions, listening before pitching, and making follow-up useful.
Our first full conversation is up:
youtube.com/watch?v=xxE8ip…
I sat down with @engstrometric to talk about asking better questions, listening before pitching, and making follow-up useful.
Our first full conversation is up:
youtube.com/watch?v=xxE8ip…
0
0
2
242
Dots in Codex/ChatGPT all of sudden asking to do everything in chat is ANNOYING! I already have approvals set in Codex/ChatGPT, why would I need separate ones in Dot...I spent a 25 minute drive today arguing with it telling it I wasn't going to click the approve button because I'm driving and am hands free...for it to ask me after the next task, "Can you click approve? I just added the task in our chat" No Franklin, I can't...
0
0
1
58
I know I need to make content...as a solopreneur, there are only so many hours and so many things I can do...this just allowed me to justify the time...
Goodbye Davinci, goodbye Premiere. We don't need you anymore... You should be editing entirely in Codex or Claude Code.
Here's exactly how to do it.
2
0
3
89
Anyone Else having codex issues like crazy today? I keep having chats freeze, start a chat and it takes forever to kick off.
Restarted codex and the computer. No fix.
I keep getting "Workspace routing discover timed out" That's a new error to me.
Restarted codex and the computer. No fix.
I keep getting "Workspace routing discover timed out" That's a new error to me.
0
0
2
113
I don't think people realize how big of a change this is. Previously, you had to use ChatGPT app using ChatGPT work or Codex with computer use...
Now, the chrome extension alone allows your chatGPT tab to understand everything else going on in chrome if you ask it to connect.
I repeat, this is NOT computer use like we've known it, it's Chrome use through an extension. Gamechanger for those that don't use the ChatGPT app.
Now, the chrome extension alone allows your chatGPT tab to understand everything else going on in chrome if you ask it to connect.
I repeat, this is NOT computer use like we've known it, it's Chrome use through an extension. Gamechanger for those that don't use the ChatGPT app.
A few big upgrades for Chrome users today! We’re bringing the power of computer use to anyone who uses Chrome.
First up: ChatGPT can now use your tabs in Chrome to make answers more useful and get things done for you.
If you’re in Chat, it can look across your open tabs and e in Work, ChatGPT can control your Chrome tabs to get things, like sending messages, filling out forms, or downloading files. It can even help you organize and group that pile of tabs you have open.
0
0
1
48
Anyone else's OpenAI Dot basically useless recently?
I keep sending feedback. I've had a handful of calls where I talk to it to get stuff done, it just says I'm checking on that or let me check on that and then NEVER ANSWERS. Not sure what the deal is there.
I keep sending feedback. I've had a handful of calls where I talk to it to get stuff done, it just says I'm checking on that or let me check on that and then NEVER ANSWERS. Not sure what the deal is there.
2
0
1
94
Everyone is so confused about this but it's NOT confusing. Most people don't care about the demand curve...
Most people use their 15-20% of usage per day in <10 hours. Does anyone want to go back to using 4-5% every 5 hours? That means you'd have to do 3-4 of the 5 hour blocks to get the same amount of work done...
Yes, more usage is great and ideal. compute isn't computing though. Demand curves are real, I know because I've done staffing and forecasting and tracked demand curves.
OpenAI WANTS to bring back 5 hour limits.
I don't want to be 3 hours into my 5 hour limit, have a crucial task I'm working on and have to pause.
Example, I have only used 25% of my weekly usage, I'm 2 hours into my 5 hour limit. full stop, limit hit. I need to get this task done. Do I waste a weekly banked reset and only get 25% for another 2ish hours of work before hitting a limit again?
I get that 2x is great but it's not so great that I want to give up my flexibility and I'm on the 20x, now 10x plan.
@thsottiaux @Dimillian @jxnlco @OpenAI @sama @ajambrosino
Most people use their 15-20% of usage per day in <10 hours. Does anyone want to go back to using 4-5% every 5 hours? That means you'd have to do 3-4 of the 5 hour blocks to get the same amount of work done...
Yes, more usage is great and ideal. compute isn't computing though. Demand curves are real, I know because I've done staffing and forecasting and tracked demand curves.
OpenAI WANTS to bring back 5 hour limits.
I don't want to be 3 hours into my 5 hour limit, have a crucial task I'm working on and have to pause.
Example, I have only used 25% of my weekly usage, I'm 2 hours into my 5 hour limit. full stop, limit hit. I need to get this task done. Do I waste a weekly banked reset and only get 25% for another 2ish hours of work before hitting a limit again?
I get that 2x is great but it's not so great that I want to give up my flexibility and I'm on the 20x, now 10x plan.
@thsottiaux @Dimillian @jxnlco @OpenAI @sama @ajambrosino
The results of this poll has gotten me confused. I definitely thought it would sway much more favorably in the 2X direction across Pro 100/200/500 at the cost of bringing back 5h limits in order to smoothen the demand curve.
0
0
1
240
Y'all, First 48-72 hours of OpenAIs dot were amazing. Last 48 hours...Frustration. Ask it a question, it just doesn't answer.
I asked it to read an email which it has the plugin for, summarize the 11 page document and let me know. I'm currently up to 7 minutes waiting.
I asked it to read an email which it has the plugin for, summarize the 11 page document and let me know. I'm currently up to 7 minutes waiting.
0
0
1
50
There are so many "model is at capacity" notices in Codex lately. Are they getting ready to release some response to Claude's new models?!?
Or are we just struggling with compute?
Or are we just struggling with compute?
0
0
1
81
OpenAI and Tibo are looking to continue building back some report with Devs after Dev day fell mostly for regular users it seems.
Over the next 28 days, each day we’ll either ship one thing that is a clear improvement and relevant for most codex/work users or ship a full reset. Let the improvements begin.
0
0
0
63
Anyone that's using OpenAI's Dot...Does it seem to be going EXTREMELY slow this morning?
The last couple days I was able to get a lot done in a 30 minute period. This morning, I feel like every time I talk to it, I get a "checking", "let me look into that" and it'll be minutes before I get any responses.
@thsottiaux @Dimillian
The last couple days I was able to get a lot done in a 30 minute period. This morning, I feel like every time I talk to it, I get a "checking", "let me look into that" and it'll be minutes before I get any responses.
@thsottiaux @Dimillian
0
0
0
30
As I sit here and work with Franklin, my Dot. I'm noticing one reason cloud, at least at first, is tough.
All my context is broken up into projects and each project has the relevant tasks and information. When it kicks tasks off in the cloud, those contextual pieces seem to be missing.
Yes, it can go find information if it knows where to look and what it may look like. But that's not always clear.
Each project, and I have MANY, has different structure and form depending on the task or agent it's built around. It feels like Dots should be able to tie cloud work to local work...That would help cleanup some of the context issues I'm seeing when trying to convert from local to cloud.
This makes me wonder, we are so used to living on our local devices as builders but many don't even have a computer and using cloud is the best way. That brings in the issue of pulling context from a device as well.
There are still some real wins in this to help make a personal context driven AI. When it comes to building and intelligence, the leader will keep shifting. I don't know that any company has figured out the personal AI thing yet and I don't know that it'll be the leader in intelligence. It'll likely be the one that can create the most efficiencies and the quickest back and forth.
This means that every time I ask a question to Franklin, I don't get, "Let me look into that for you". Ultra Fast and efficient models that need to reason less for the same output are the answer here.
Each company is going for a different goal line, many don't see it because they focus on the revenue and business side of AI. It feels like the AI companies are starting to differentiate their goals while still trying to be relevant to one another and prevent one company from running away in a new product market.
Do you have any insight here? Anything I may have missed?
All my context is broken up into projects and each project has the relevant tasks and information. When it kicks tasks off in the cloud, those contextual pieces seem to be missing.
Yes, it can go find information if it knows where to look and what it may look like. But that's not always clear.
Each project, and I have MANY, has different structure and form depending on the task or agent it's built around. It feels like Dots should be able to tie cloud work to local work...That would help cleanup some of the context issues I'm seeing when trying to convert from local to cloud.
This makes me wonder, we are so used to living on our local devices as builders but many don't even have a computer and using cloud is the best way. That brings in the issue of pulling context from a device as well.
There are still some real wins in this to help make a personal context driven AI. When it comes to building and intelligence, the leader will keep shifting. I don't know that any company has figured out the personal AI thing yet and I don't know that it'll be the leader in intelligence. It'll likely be the one that can create the most efficiencies and the quickest back and forth.
This means that every time I ask a question to Franklin, I don't get, "Let me look into that for you". Ultra Fast and efficient models that need to reason less for the same output are the answer here.
Each company is going for a different goal line, many don't see it because they focus on the revenue and business side of AI. It feels like the AI companies are starting to differentiate their goals while still trying to be relevant to one another and prevent one company from running away in a new product market.
Do you have any insight here? Anything I may have missed?
0
0
2
138