Adding a feature because ChatGPT incorrectly thinks it exists

Mercury: Ultra-fast language models based on diffusion

I used o3 to profile myself from my saved Pocket links

New sphere-packing record stems from an unexpected source

Neanderthals operated prehistoric “fat factory” on German lakeshore

When Figma starts designing us

Hymn to Babylon, missing for a millennium, has been discovered

Hey HN, I’m Tejas at Morph. We’ve built a blazing-fast model for applying AI-generated code edits directly into your files at 4,500+ tokens&#x2F;sec. No more slow full-file rewrites or brittle search-and-replace hacks.Here&#x27;s a demo video: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=LdT8epGHJPk" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=LdT8epGHJPk</a>.Why? AI spits out code that can’t reliably be inserted into existing code. Full file rewrites, brittle search-and-replace hacks are too slow, expensive, or error-prone.Morph&#x27;s approach:- Your agent outputs edits “lazily”, referencing unmodified lines in the existing file (ex: &#x2F;&#x2F; ...existing code...)- Morph instantly applies these edits to a file using our Fast Apply model + speculative decoding against the original file, making AI patches fast, reliable, and production-ready.This approach was pioneered by Cursor last year, but their models aren’t available as APIs—so we built Morph for developers everywhere (with a large free tier!)Live demo (no signup): <a href="https:&#x2F;&#x2F;morphllm.com&#x2F;dashboard">https:&#x2F;&#x2F;morphllm.com&#x2F;dashboard</a> and docs: <a href="https:&#x2F;&#x2F;docs.morphllm.com&#x2F;quickstart">https:&#x2F;&#x2F;docs.morphllm.com&#x2F;quickstart</a>We have 2 Fast Apply models: morph-v3-fast - 4500+ tok&#x2F;sec, and morph-v3-large - 2500+ tok&#x2F;sec. These models power Fast Apply at create.xyz, databutton, continue.dev, and more!We also provide retrieval models for embedding + reranking.
Next Up: Inline Edit Model (Cmd-K): Extremely fast inline edits - keep dev flow state; and Morph Tab API: Our Next Edit Prediction model guesses your next code edit + action with sub-500ms latency. It&#x27;s currently in private beta, but you can request early access here: <a href="https:&#x2F;&#x2F;morphllm.com&#x2F;tab">https:&#x2F;&#x2F;morphllm.com&#x2F;tab</a>Hot takes:1) Raw inference speed matters more than incremental accuracy gains for dev UX—agree or disagree?2) Full-file rewrites by frontier models are legacy—Fast Apply edits win on speed, cost, reliability.3) As benchmarks on narrow tasks saturate to 99%+, complexity is shifting from single frontier models to specialized inference-optimized models. As frontier models move upmarket, they&#x27;ll leave simple tasks behind, and they&#x27;ll be used to do tasks only frontier models can doWe’d love to hear your ideas and experiences with coding agents!

Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

François Chollet: The Arc Prize and How We Get to AGI [video]

Solving Wordle with uv's dependency resolver

Hello HN!As a long term NYC resident, I have read countless articles on ideas tweaking subway services, but always found them hard to follow without visual aid. So over the long weekend I decided to build one. It has all the basic features: trains would spawn at their origin, stop at stations, and slow down if it gets too close to another. You can also design custom routes by piecing tracks together.Have fun, and let me know what you think!

HN

Adding a feature because ChatGPT incorrectly thinks it exists

Mercury: Ultra-fast language models based on diffusion

I used o3 to profile myself from my saved Pocket links

New sphere-packing record stems from an unexpected source

Neanderthals operated prehistoric “fat factory” on German lakeshore

When Figma starts designing us

Hymn to Babylon, missing for a millennium, has been discovered

Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

François Chollet: The Arc Prize and How We Get to AGI [video]

Solving Wordle with uv's dependency resolver

Show HN: NYC Subway Simulator and Route Designer

My first verified imperative program

Lightfastness Testing of Colored Pencils

The Miyawaki Method of micro-forestry

The chemical secrets that help keep honey fresh for so long

Running a Certificate Transparency log

The era of exploration

Why are there no good dinosaur films?

What Microchip doesn't (officially) tell you about the VSC8512

Trying to find meaning in owning an old Mac