@catalinmpit If you're writing 7-word lowercase sentences with slang, improper spelling/grammar, etc., I don't know what you'd otherwise expect.
Garbage in, garbage out.
English
Random Libertarian Tech Lead
26.3K posts
Random Libertarian Tech Lead
@someRandomDev5
Reason over feelings, whenever the two come into conflict.
Pronouns: Who/cares Bergabung Şubat 2018
67 Mengikuti422 Pengikut
@catalinmpit Seems to be a pattern:
x.com/catalinmpit/st…
Catalin@catalinmpit
A short AI story.
English
Lately, Claude makes some shocking mistakes.
⟶ Implements overly complex code
⟶ Ignores the codebase's code style
⟶ Removes working code for no reason
⟶ Replaces code that's out of scope from the task at hand
It feels like it needs 100% supervision. At this point, you're better off writing everything yourself.
English
Think LLMs are unreliable?
Maybe it's not the model. Maybe it's you.
Most people blame the AI when their outputs fall apart. But they never audit their own prompts. They don't check if they gave enough context. They expect it to read their mind.
You want better results? Start with better input. LLMs reflect your clarity. If you're vague, they will be too.
The model isn't failing. Your process is.
English
@Yampeleg There's nothing inherently wrong with doing it.
What people are mad about is that cursor has implied this whole time that they built Composer from the ground up.
English
@samtwtss The funniest thing about AI is that it's always good at creating what you ask for... as long as what you ask for isn't that creative.
The more worthless it is to the world, the better a job AI does at making it. 😂
English
bro, it's so over for designers
google stitch is insane. 🤯
Stitch by Google@stitchbygoogle
Meet the new Stitch, your vibe design partner. Here are 5 major upgrades to help you create, iterate and collaborate: 🎨 AI-Native Canvas 🧠 Smarter Design Agent 🎙️ Voice ⚡️ Instant Prototypes 📐 Design Systems and DESIGN.md Rolling out now. Details and product walkthrough video in 🧵
English
English
@someRandomDev5 @fynnso yeah good point, what form of fine tuning you think they ended up using?
English
@FelixKerlin @cryptopunk7213 Savings per task would be a concise way to express it but then it would not be immediately clear that it's relative to only the most expensive model on this chart (rather than being some arbitrary metric, or relative to the most expensive model in existence).
English
@FelixKerlin @cryptopunk7213 Since you already have a ceiling value that's constrained to this specific set of models, you could just frame it as "cost savings relative to the most expensive"... But that's difficult to communicate concisely in that amount of space.
English
1. theres no way this beats opus 4.6
2. wtf is this chart crime??? 😂
Cursor@cursor_ai
Composer 2 is now available in Cursor.
English
@wesbos ... And there are some leaks which seem to suggest that Cursor's Composer model is no different; it seems that it may just be a fine-tuned version of Kimi K 2.5:
Fynn@fynnso
was messing with the OpenAI base URL in Cursor and caught this accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast so composer 2 is just Kimi K2.5 with RL at least rename the model ID
English
@wesbos Opus and GPT 5.4 are both highly intelligent.
But Opus has the edge in understanding people, and GPT 5.4 has the edge in understanding systems.
Nearly every open-weights model is basically just a distilled fine-tuned Frankenstein of GPT, Claude, and Gemini.
English
Composer 2 vs Opus 4.6 vs GPT 5.4 - a totally unscientific test
> Create a Twitter clone. Use Better Auth, Vite, Sqlite, Drizzle, Typescript, and React with Tanstack Start.
Each one took ~5 mins in plan mode. Each had access to a browser to test.
Composer: 5 mins $6.04 1,250 LOC
Opus: 19 mins $10.43 1,000 LOC
GPT: 22 mins $14.15 2,000 LOC
Opus seems to use the cache WAY more than composer, so it's not really 10x more expensive.
Composer app ran first try. Other two needed a bit of CORS debugging but did work.
Code between all was extremely similar
All three done inside cursor - so Claude Code / Codex may have been different
Models used:
Composer 2.0 (regular, not fast)
Opus 4.6 Medium Thinking
GPT 5.4 Medium Thinking
Wes Bos@wesbos
Cursor just launched Composer 2 - their own model. It's 10× cheaper than Opus 4.6 and supposed to rival it. I've been using it for a few days, I don't have any skewed graphs to show you but from a pure vibes POV I can tell you it's pretty good™ My litmus test right now is if it can build a 3D Printable model with Manifold CAD. I build a Gif zoetrope generator and it did fantastic.
English
@wesbos If you want an interesting test that actually demonstrates brownfield development (without requiring too much set up on your own part), ask the model to expand or modify the behavior of a large complicated open-source project that you actively use.
English
@wesbos That said, this greenfield test does do a fairly good job of demonstrating how well these models can replace design agency work. If all you're building are businesscard-like or brochure-like landing pages, then sure, this test is fine.
English
English
@fynnso In this scenario is it actually fine tuning or is it a LoRa adapter?
I’m a little new to these concepts but very curious.
English
was messing with the OpenAI base URL in Cursor and caught this
accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast
so composer 2 is just Kimi K2.5 with RL
at least rename the model ID
Cursor@cursor_ai
Composer 2 is now available in Cursor.
English
@BlackHC TLDR: How did you know that software was good before the era of AI? Pure vibes. How do you know it's good after the era of AI? Still pure vibes, nothing has changed.
English
@BlackHC Again, prompt engineering is an actual skill set and there's actually a lot of nuance to it despite people ridiculing it.
"Building an intuitive sense of the things that the LLM can actually infer accurately through language" is a key part of that.
English
A while back, Andrej Karpathy said the app store will be replaced by generated, disposable software," and Amjad Masad predicted that the value of all application software will go to zero
I think this "ephemeral software hypothesis" is wrong, though, and I want to explain why:
English