polymorph3us's profile picture. junior dev | happy to be here | LOTR | functions | types

polymorpheus

@polymorph3us

junior dev | happy to be here | LOTR | functions | types

polymorpheus 已轉發

Holy... You're telling me Gemini 3 made this? Which would take a human UI/UX and frontend engineer weeks of work…

來自 Cali Castle

polymorpheus 已轉發

Switched Context7 from token-based to page-based retrieval — here’s what changed: ✅ 35% fewer tokens (5.4K → 3.5K) ⚡ 19% faster responses 🎯 Accuracy maintained at 7/10 630 questions benchmarked. Same quality — dramatically better efficiency. Results below.

Context7AI's tweet image. Switched Context7 from token-based to page-based retrieval — here’s what changed:

✅ 35% fewer tokens (5.4K → 3.5K)
⚡ 19% faster responses
🎯 Accuracy maintained at 7/10

630 questions benchmarked. Same quality — dramatically better efficiency. Results below.

polymorpheus 已轉發

Amazing test of Gemini 3’s multimodal reasoning capabilities: try generating a threejs voxel art scene using only an image as input Prompt: I have provided an image. Code a beautiful voxel art scene inspired by this image. Write threejs code as a single-page


polymorpheus 已轉發

We’ve been putting Gemini 3 Pro through real design tests in Figma Make: holiday projects, extreme aesthetic swings, and system-level tasks. It handled the fundamentals reliably and showed meaningful range in layout, style, and interaction. The pattern is clear: AI widens the…


polymorpheus 已轉發

coffee brew calculator by gemini 3


polymorpheus 已轉發

Introducing the Parallel FindAll API. Create an entire dataset with a single search query. For example: "Find all dental practices in Ohio with a 4+ star review on Google" parallel.ai/blog/introduci…


polymorpheus 已轉發

the question I get the most at every conference or meetup: "what even are durable objects?" I spent the past 2 years building almost exclusively with durable objects, and today I'm compiling my notes and favourite patterns into a single blog post boristane.com/blog/what-are-…

boristane's tweet image. the question I get the most at every conference or meetup: "what even are durable objects?"

I spent the past 2 years building almost exclusively with durable objects, and today I'm compiling my notes and favourite patterns into a single blog post

boristane.com/blog/what-are-…

polymorpheus 已轉發

Effect + DOs is going to be a blast. We're focused on this @alchemy_run with alchemy-effect, but aim to take it even further with universal, cross-cloud, type-safe bindings. Cloudflare at the root is perfect. In their words: "the Connectivity Cloud"

i hope yall are ready for me to be insufferable on the tl about durable objects



polymorpheus 已轉發

The most useful Category Theory abstractions: 1. Functor 2. Monad 3. Applicative 4. Semigroup 5. Monoid Honestly, if we can make at least these ideas mainstream, that would be good enough. Other common useful abstractions: 1. Contravariant 2. Profunctor 3. Bifunctor 4. Lenses…


polymorpheus 已轉發

I used Monoids, Applicative Functors, Contravariant and Comonads to design a composable logging library. I used Profunctors, Isomorphisms and Categories to design a correct-by-construction decoding library. There's so much power in leveraging foundational abstractions.


polymorpheus 已轉發

Currently using the best of everything: Package Manager → @bunjavascript Frontend → @tan_stack Start Authentication → @better_auth Backend → @convex Blob storage → @Cloudflare R2 Payments → @polar_sh / @autumnpricing Email → @resend Observability → @Sentry +…


polymorpheus 已轉發

Introducing group chat on Poe: a new way to collaborate with any AI and anyone you know, all in a single conversation. Now available to all users worldwide, groups can interact with any of the 200+ text, image, video, or audio models on Poe, plus any creator-made bots. (1/8)


polymorpheus 已轉發

Open sourcing InScene + InScene Annotate - a pair of LoRAs for steering in-scene image generation w/ QwenEdit. You can use a combination of prompts + annotations to generate consistent images inside a scene. In beta but v. powerful. Model + workflow below, training data soon.


polymorpheus 已轉發

"I am a full-on Cloudflare Zero Trust with Warp convert, and while I still have Tailscale running in parallel, almost everything I do now is going through Zero Trust tunnels." Great writeup on Cloudflare Tunnels from @dvcrn! david.coffee/cloudflare-zer…


polymorpheus 已轉發

when people diagnose “lack of agency” the real issue is lack of playfulness. fix playfulness and agency follows naturally. trying to "will" agency into your life as if you’re somehow broken only perpetuates the problem


polymorpheus 已轉發

my only advice to people who ask me for advice: you must learn to tell simple compelling stories. this skill will compound like crazy in every facet of your life. that’s it.


polymorpheus 已轉發

getting in just under the wire happy to present offworld.sh as my submission to the @tan_stack @convex hackathon 🫡 search for any public github repo to view its AI generated analysis (or kick one off!), including overall project summary, architecture details…


polymorpheus 已轉發

Debugging CI/CD pipelines locally was too complex & expensive to solve. AI tools changed that. We built Magnolia—run GitHub Actions, GitLab CI, and Forgejo pipelines locally. github.com/tuist/magnolia Not feature-complete yet, but open to contributions. No more push-pray-wait.


polymorpheus 已轉發

Kiro is generally available 👻 Specs made ‘planning first’ the default for AI assisted dev. Now Kiro IDE adds property based tests to check if your code actually matches your Spec. Real signals, not vibes. Plus a new Kiro CLI and full team support through AWS IAM Identity…


polymorpheus 已轉發

supervision-0.27.0 lets you parse and visualize Qwen3-VL object detection results prompt: person between albert and marie. to answer this, Qwen3 needs prior visual knowledge of Marie Curie and Albert Einstein, and it needs to understand reference terms like between.

skalskip92's tweet image. supervision-0.27.0 lets you parse and visualize Qwen3-VL object detection results

prompt: person between albert and marie.

to answer this, Qwen3 needs prior visual knowledge of Marie Curie and Albert Einstein, and it needs to understand reference terms like between.

how many taxis do you see in this image? Qwen3-VL is so good at recognition, multi-target grounding, and understands spatial relations. it blows my mind.

skalskip92's tweet image. how many taxis do you see in this image?

Qwen3-VL is so good at recognition, multi-target grounding, and understands spatial relations. it blows my mind.


United States 趨勢

Loading...

Something went wrong.


Something went wrong.