Pinned Tweet
Have been giving GPT 6 Astra a genuinely good two days(mostly because I maxed out my ant sub)
first of all I don't remember Astra being this good to work with but on the other hand Astra was always hugely expensive so I guess I never gave it a real shot
now with banked resets and such working with it on High effort yeah this is awesome
secondly been working on how to display in-transcript images better and also how to have them show in a sidepane when in-transcript doesn't suffice but also on a canvas based on tldraw with all that gives
after 1 hour of ideation or such it produced a very nice concept off the bat
65
suuuuper rough and basically a first draft after brainstorming and ideation with Fable 5.2 high
after about 10 hours of work by fable 5.2 on medium this is the result. It captures almost everything we agreed on and I never really gave much in terms of the design but rather what I wanted and needed from this
Didn't know what to expect but this is actually usable, it's not pretty and it's buggy but it works. I could make an iphone app with this and test it on my phone, from my custom tooling.
this is every day things now but still this is nuts
1
1
16
been meaning for the longest time to build support for creating iphone apps in my custom tooling but always had more important or fun things in line
now though I have a project for someone close to me that I wanna get going so it's time to put this feature at the top of the pile
one caveat, I don't have a mac I'm on windows but me and them use iphones.
once the phone actually got recognized on windows 10 it was very quick from there for fable 5.2 to actually get something going.
this is for it's internal use, just to see the fundamentals are in place before building anything or before even brainstorming what this should be and such but it took like 20 minutes from "it's recognized" to this showing up in the transcript
impressive actually
1
1
31
I'm starting a new project today - a multi-year quest to save my hand from being completely immobilised.
In 2019 i broke my left wrist 2 times within a month. First time happened after i crashed on a bicycle - which ended up in a surgery, metal plate and multiple bolts in my bones. A month later another accident made this bolts work as weak points, completely destroying the bones from the inside.
After my second surgery, the surgeon told me I bought around 15 years - after which the pain will become so unbearable I'll need to replace the bones with metal rods, removing all the mobility in my wrist.
So, my goal is to solve this problem within the next 7 years. The first phase of this project is to gather the data: i'm trying to get high quality 3d scans of every possible layer - soft and hard tissues in my wrist.
I've started with the 3D CT scan to map the bone structure. The website compares both of my wrists to make it easier to understand the current situation.
What I'm looking to do next:
• get more high quality scans (help me by recommending clinics/hospitals with good equipment)
• get more opinions and ideas (help me by sending this to orthopedic surgeons you know)
• get more knowledge on the topic (help me by pointing what exactly should i read/learn/watch on bioengineering).
not sure if this thing is even solvable, but we gotta try huh
wrist.kyzo.io
Fable 5.2 has been hill climbing how to make the composer for my custom tooling the smoothest and fastest typing experience ever to exist....or close enough
it's been going for over 3 hours now
the prompt(yes I always say hello actually):
hello
I want to make the composer in session projection pane the most smooth experience for typing that ever existed. I want it to feel like the characters appear before I hit the keyboard(figuratively). Can you help me achieve this?
1
1
27
after working with the previous version for a few hours there were some things I didn't like:
- the column of buttons in the middle
- alignment was still off, especially when expanding/collapsing lists
- paths was not highlighted as clickable interactions
- indentation was hardly noticeable(yes there was some but not enough, at all)
and after dogfooding that for....2 hours I noticed more things(second picture)
quite pleased with the results
now the only thing that brings greater returns is to just use it and see where it breaks and such
1
1
31
quite pleased with the results
now the only thing that brings greater returns is to just use it and see where it breaks and such
working on a tasks and projects tool
first iteration by claude looked great but wait a minute....
design really is such a nit picky thing and it never really feels complete. In many ways just like software dev.
Seems like Google actually cooked at last
Been giving Fable 5.2 everything I can throw at it and it just makes it happen, in some cases where opus 5.5 struggled a bit.
Now Carbon is on that level? damn
While being 4x cheaper(on average)? damn
What a day! I tested Gemini 4 Carbon on real work. 2 repos, 105 replanted bugs. Find and fix what you can.
It scores just behind Fable 5.2. The results:
- Fable 5.2 (max): 58 for $139.65
- Gemini 4 Carbon (high): 57.3 for $31.32
- Sonnet 5.5 (max): 51.3 for $154.03
- GPT-6 Astra (max): 45 for $33.03
- GPT-6.1 Sol (max): 44.3 for $5.88
- Gemini 4 Argon (high): 40.7 for $7.16
Google's own benchmarks looked super strong, but nobody could access the model. Turns out they didn't lie: +16.6 points over Argon, at less than a quarter of Fable 5.2's cost.
Nobody should be laughing at Google anymore.
(the next effort level drops soon in the thread 🧵)
1
2
113
Are you possibly being reouted to GPT 6.2 Sol? ask gpt 6.1 Sol about Fable 5, tell it to not use memory or web search
If it knows about Fable 5 then congrats you are being routed!
I am not being routed to 6.2 Sol :(
I never thought I'd live to see GPT-6.2 Sol being SO MUCH BETTER than GPT-6 Astra while being significantly cheaper. This is absolutely incredible. OpenAI is back.
This entire village was built with Three.js, packed into a single HTML file. No Blender, no external 3D assets. Just code.
The cherry blossoms, the river, the whole scene shifting from daylight to glowing windows at night. Look at this.
1
1
95
Been working on a pull request tool, or perhaps environment even?, for use at work where we do a lot of code review and after literally 4 hours of ideation with claude we got a rough.....very rough first draft
so much slop but also so much potential and it basically got all the primitives and such into place
but man it was such a mess, I've captured some of it in the first two pictures(this is all palceholder data)
gave a lot....really a lot of feedback on what I liked and what had to change or why it couldn't remain like that
also asked it to use @heydamir unslop skill github.com/heydamir/unslop
already after one iteration it's basically like two different dimensions(second set of screenshots)
still some rough edges like the dual scroll handles but this is in a small window inside a side pane in claude code so that is definitely not gonna make it into my custom tooling
2
1
72
Generally I also like to have layouts and generally UI and UX that isn't techy or saas. It's still a technical tool but I try(emphasis on try) not to have it always look like a tech tool
so I asked claude to give me a sort of inbetween
again some rough edges but it's not meant to be perfect since this is just still demo and look and feel
1
44
anthropic has done a great job updating claude code desktop app to feel more performant and like an actual tool and not just a chat interface with panes bolted on
very nice
not good enough to replace my custom tooling but overall this is a much improved experience
1
1
45



























