It's interesting to consider what happening here. Clearly they were on a time crunch. So either they used GPT 5 to do it, and it messed up this badly. Or they trust GPT 5 so little internally that they did it manually. Both are not good looks…
GENERATIVE AI
-

GPT-5 Ranks First on Multiple LM Arena Benchmarks
By
–
GPT-5 is already 1 on @lmarena_ai in Text, WebDev, and Vision Arena in Hard Prompts, Coding, Math, Creativity, Long Queries, and more https://
x.com/lmarena_ai/sta
tus/1953504958378356941
… -

GPT-5 Ranks #1 Overall on LM Arena After Blind Testing
By
–
GPT-5 #1 toute catégories confondu sur LM ARENA 1 million d'utilisateurs l'ont testé sans savoir que c'était GPT-5, sous le nom de code "summit"
-
GPT-5 Creates Impressive Poem Camera App With Auto-Generated Details
By
–
A compilation of experiences I made with GPT-5 in one shot.
— Pietro Schirano (@skirano) 7 août 2025
The poem camera app is particularly impressive because the model came up with all the details, like the way the photos stack in the gallery, the photo developing animation, etc pic.twitter.com/MHHAMVDdcBA compilation of experiences I made with GPT-5 in one shot. The poem camera app is particularly impressive because the model came up with all the details, like the way the photos stack in the gallery, the photo developing animation, etc
-
GPT-5 Prompts Sharing Initiative Announced
By
–
Thank you! Will definitely be posting GPT-5 prompts there 🙂
-

GPT-5 Matches Gemini Pricing with Better Coding Performance
By
–
GPT-5 is priced at the same level as Gemini, appears to be slightly better than Gemini (for coding at least). That's some decent progress, although I think a lot of folks were hoping for more.
(h/t @simonw for the table) -
AI Tool Generates Custom Design Styles From Screenshots
By
–
You can ask it for different design styles! You can even give it a screenshot of a design you like and it can take inspiration from that!
-
Build Complex Software Beyond Simple SaaS Projects
By
–
Try going beyond simple SaaS builds. What is a piece of software you've wanted to build but previously would take months w/ a great, specialized team? Try that 🙂
-
Evaluating GPT-5: Testing AI Beyond Current Capabilities
By
–
For those who will be using/testing GPT-5: You won't see much improvement if you're asking it to do things Sonnet/o3 can already do. To see how good it truly is, you have to ask it to do things that other AIs simply can't.
