raphh (@raphh) header
raphh (@raphh) profile picture

Product Engineer. iOS/macOS & more. Prev @Immoweb, @SeLoger, @M6 and @Bedrock_Stream

Brussels
Joined December 2008
"Take a step back" never disappoints.
15
Over the years I've been doing most of my movie watching at home, but when there's an IMAX screening it's usually my cue to drag my ass back to the theater, and I never regret it. These are such humbling experiences that push cinema to another level. Feeling blessed to be able to appreciate this kind of art and to have an IMAX 70mm theater in my city. So far I've only been going for Nolan movies, since he's obviously the master at it, but I booked my ticket for Dune: Part Three too after seeing that insane 4-5 minute sequence before The Odyssey. Some movies are really meant to be experienced in a theater, and I'll try to honor that more going forward.
1
71
Back with GPT 6.1-Sol bullshit for a bit, didn't miss it at all lol
36
First weekly limit hit on Max 20x. Taking it easy until Sunday I guess 馃ゲ
23
Other than that, Projects are really great. I love how it can switch from the project's remote machine to my machine to my server, all in one thread. Can be a bit iffy sometimes, but overall it's going in the right direction.
1
19
I feel like Claude gets stuck on this too often. Would love a setting in each Project to allow this for every thread in it. Not really sure I get the logic of how it works right now.
20
I asked Claude to help me build my own version of @poteto's pstack. One thing I'm adding is the trust ladder concept, plus mechanisms that help agents find opportunities to "close the loop".
44
Better late than never: I finally got a ticket for The Odyssey in 70mm IMAX at Kinepolis Brussels. They added a bunch of screenings for October, and this is the best seat I could find (at 9:45am though lol)
1
41
Opus and Sol had a productive night. Not at @poteto level yet, but it's a start. For review, Claude Artifacts are really goated. Opus creates an explainer artifact for each PR, plus a mega explainer showing what happened across the whole run. Once I've reviewed it and sent feedback, I usually ask my agents to re-consolidate the PR stack in the way that makes the most sense given my feedback. They re-read the new PRs, then I merge. Then I spend the rest of the day "gardening": reading through the codebase and finding ways to tighten the environment so my agents can produce more efficient runs next time. At some point I decide the next milestones and prepare a run plan for my agents to work on overnight, and repeat. What takes the most time is the "assimilation phase": understanding what happened during the run and what the agents did, both in product and in process. It's the only way to course correct and make our runs better over time.
37
Claude just gave me a thumbs-up lol. Also the desktop app is very good now, good work from the people at @ClaudeDevs.
1
20
We are still on the x20 limits right now. We'll have to wait for October 30 to see what the new Pro 200 limits sounds like.
Replying to @Dimillian
6.1 Sol is very good. Been running it all day on xhigh/max with multiple threads and it barely moved the weekly limits.
47
Free credits! I don't think I've seen any mention of this from anyone at OpenAI? Curious to see how often I'll go over limits with GPT-6.1 Sol and the change in the Pro200 plan.
53
Replying to @sama
EU when Sam? 馃ゲ
10
1
128
19,932
GPT-6.1 Sol is already available in Codex.
42
GPT-5.6 Sol, Luna and Terra will be available as unlimited with the ChatGPT Pro subscription. This is already visible on the site: chatgpt.com/pricing/
GPT-5.6 Sol, Luna and Terra are listed as "Unlimited*" usage on the Pro plan page. It looks like OAI are giving unlimited (*with fair usage) use of those models to balance out this change in pricing. 5.6 Sol and Luna X-High are two of my fave models so this could be great!
98
I asked Codex to explain this post to me. It can sound a bit scary for sure, and that's probably why Tibo announced it before the event. Let's see how it plays out in real life. Pretty sure they'll monitor it and adjust if it goes wrong.
Hi, Tomorrow we are re-opening the Pro $200 subscriptions to new subscribers, but together with it we are also changing how we calculate the usage for it. In effect, if you do the math, it will net out at half the dollar in API spend compared to the old Pro $200 plan. Now that it's said, let me explain why this is happening and why you will still get more work done than if you were on the Pro $200 subscription one month ago. (a) We didn't want to compromise in other ways and are committing to not reintroducing the 5h limit, so that you can fully use the weekly usage when you want. (b) On the subscription, we guarantee that over time you always get more work done and with an increasing level of quality. This means that you will continue to get more value per dollar spent as a result of models getting more efficient and us passing down the improvements in the form of API price reductions. (c) We don't want to put an incentive on ourselves to artificially inflate the API list prices to make it look like you are getting a lot (and workaround it through discounts, etc). Instead we want to continue to both rapidly reduce prices and increase capabilities of models on the API. This week we introduced GPT-6 Sol and GPT-6 Luna at 50% of their previous price. Over time, we see prices go low enough that it makes sense for most to buy usage as needed without there being a significant gap between what you get in a subscription and what you get in the API for a dollar spent. (d) Tomorrow, we are adding more things to the subscription that won't draw on the usage, I won't reveal what that is yet. I wanted to be transparent before all the big announcements tomorrow. Lots of new exciting things are coming to the subscriptions that will make it super compelling, but I wanted to make sure to share this change ahead of time so you can all understand it before we shower you with good news. Codexingly, Tibo
30