The best Hacker News stories from Show from the past day

from

All Show

from the past*

day week month

with these search terms:

Create follow Go back

Latest posts:

Show HN: Torch Lens Maker – Differentiable Geometric Optics in PyTorch

Hello HN! For the past 6 months I've been working on an open source python library that implements differentiable geometric optics in PyTorch. It's very experimental still, but eventually the goal is to use it to design optical systems with a state of the art optimization framework and a beautiful code based API. Think OpenSCAD, but for optical systems.<p>Not only is PyTorch's autograd an amazing general purpose optimizer, but torch.nn (the neural network building blocks) can be used pretty much out of the box to model an optical system. This is because there is a strong analogy to be made between layers of a neural network, and optical elements in a so-called sequential optical system. So the magic is that we can stack lenses as if we were stacking Conv2D and ReLu layers and everything works out. Instead of Conv2D you have ray-surface collision detection, instead of ReLu you have the law of refraction. Designing lenses is surprisingly like training a neural network.<p>Check out the docs for examples of using the API. My favorite one is the rainbow :) <a href="https://victorpoughon.github.io/torchlensmaker/examples/rainbow" rel="nofollow">https://victorpoughon.github.io/torchlensmaker/examples/rain...</a><p>You should be able to `pip install torchlensmaker` to try it out, but I just set it up so let me know if there's any trouble.<p>I was part of the Winter 1'24 batch at the Recurse Center (<a href="https://www.recurse.com/" rel="nofollow">https://www.recurse.com/</a>) working on this project pretty much full time. I'm happy to talk about that experience too!

Show HN: Torch Lens Maker – Differentiable Geometric Optics in PyTorch

Hello HN! For the past 6 months I've been working on an open source python library that implements differentiable geometric optics in PyTorch. It's very experimental still, but eventually the goal is to use it to design optical systems with a state of the art optimization framework and a beautiful code based API. Think OpenSCAD, but for optical systems.<p>Not only is PyTorch's autograd an amazing general purpose optimizer, but torch.nn (the neural network building blocks) can be used pretty much out of the box to model an optical system. This is because there is a strong analogy to be made between layers of a neural network, and optical elements in a so-called sequential optical system. So the magic is that we can stack lenses as if we were stacking Conv2D and ReLu layers and everything works out. Instead of Conv2D you have ray-surface collision detection, instead of ReLu you have the law of refraction. Designing lenses is surprisingly like training a neural network.<p>Check out the docs for examples of using the API. My favorite one is the rainbow :) <a href="https://victorpoughon.github.io/torchlensmaker/examples/rainbow" rel="nofollow">https://victorpoughon.github.io/torchlensmaker/examples/rain...</a><p>You should be able to `pip install torchlensmaker` to try it out, but I just set it up so let me know if there's any trouble.<p>I was part of the Winter 1'24 batch at the Recurse Center (<a href="https://www.recurse.com/" rel="nofollow">https://www.recurse.com/</a>) working on this project pretty much full time. I'm happy to talk about that experience too!

Show HN: A terminal emulator in pure PHP

Show HN: A terminal emulator in pure PHP

Show HN: A terminal emulator in pure PHP

Show HN: A terminal emulator in pure PHP

Show HN: Cursor Directory – From a 3-hour build to a 250k users/mo community

Viktor and I built the first version of Cursor Directory in just 3 hours, right after I watched a few downloaded YouTube videos on my flight to France. The idea was simple: create a single place to find curated rules to enhance the Cursor experience.<p>Fast forward a few months, and the site kept growing. We recently added support for MCPs, a Trending board, the ability to generate your own rules from dependencies, and more!<p>We’d love to hear your thoughts—let us know if you have any ideas to make Cursor Directory even better!<p>Oh, and we’re fully open source!

Show HN: Minimalytics – a standalone minimal analytics app built on SQLite

Hi everyone! I wanted to share my analytics app with you.<p>This project came from requirements to track certain very frequent events. I found that the cost to do it on a regular analytics product was much more than i was willing to pay. Secondly, I also wanted to use as few resources as possible. So I thought it may be a good idea to create something that may be useful for myself (and hopefully others).<p>I have been able to track a great number of events with this using ~20 MB of storage and memory which is incredible. I have been really impressed by golang as a language and as an ecosystem and would love to work more in this language going forward.<p>Some Highlights: 1. No dependencies 2. CLI based management 3. Web based UI (and the server to serve it) included in the program. 4. 20 MB install size. 5. 20 MB memory use while running. 6. Minimal storage requirements because it aggregates events.<p>This can be a great fit for anyone who wants to have a lightweight minimal analytics for internal events.<p>I am looking forward to your comments and feedback.

Show HN: Minimalytics – a standalone minimal analytics app built on SQLite

Hi everyone! I wanted to share my analytics app with you.<p>This project came from requirements to track certain very frequent events. I found that the cost to do it on a regular analytics product was much more than i was willing to pay. Secondly, I also wanted to use as few resources as possible. So I thought it may be a good idea to create something that may be useful for myself (and hopefully others).<p>I have been able to track a great number of events with this using ~20 MB of storage and memory which is incredible. I have been really impressed by golang as a language and as an ecosystem and would love to work more in this language going forward.<p>Some Highlights: 1. No dependencies 2. CLI based management 3. Web based UI (and the server to serve it) included in the program. 4. 20 MB install size. 5. 20 MB memory use while running. 6. Minimal storage requirements because it aggregates events.<p>This can be a great fit for anyone who wants to have a lightweight minimal analytics for internal events.<p>I am looking forward to your comments and feedback.

Show HN: Minimalytics – a standalone minimal analytics app built on SQLite

Hi everyone! I wanted to share my analytics app with you.<p>This project came from requirements to track certain very frequent events. I found that the cost to do it on a regular analytics product was much more than i was willing to pay. Secondly, I also wanted to use as few resources as possible. So I thought it may be a good idea to create something that may be useful for myself (and hopefully others).<p>I have been able to track a great number of events with this using ~20 MB of storage and memory which is incredible. I have been really impressed by golang as a language and as an ecosystem and would love to work more in this language going forward.<p>Some Highlights: 1. No dependencies 2. CLI based management 3. Web based UI (and the server to serve it) included in the program. 4. 20 MB install size. 5. 20 MB memory use while running. 6. Minimal storage requirements because it aggregates events.<p>This can be a great fit for anyone who wants to have a lightweight minimal analytics for internal events.<p>I am looking forward to your comments and feedback.

Show HN: AgentKit – JavaScript Alternative to OpenAI Agents SDK with Native MCP

Hi HN! I’m Tony, co-founder of Inngest. I wanted to share AgentKit, our Typescript multi-agent library we’ve been cooking and testing with some early users in prod for months.<p>Although OpenAI’s Agents SDK has been launched since, we think an Agent framework should offer more deterministic and flexible routing, work with multiple model providers, embrace MCP (for rich tooling), and support the unstoppable and growing community of TypeScript AI developers by enabling a smooth transition to production use cases.<p>This is why we are building AgentKit, and we’re really excited about it for a few reasons:<p>Firstly, it’s simple. We embrace KISS principles brought by Anthropic and HuggingFace by allowing you to gradually add autonomy to your AgentKit program using primitives:<p>- Agents: LLM calls that can be combined with prompts, tools, and MCP native support.<p>- Networks: a simple way to get Agents to collaborate with a shared State, including handoff.<p>- State: combines conversation history with a fully typed state machine, used in routing.<p>- Routers: where the autonomy lives, from code-based to LLM-based (ex: ReAct) orchestration<p>The routers are where the magic happens, and allow you to build deterministic, reliable, testable agents.<p>AgentKit routing works as follows: the network calls itself in a loop, inspecting the State to determine which agents to call next using a router. The returned agent runs, then optionally updates state data using its tools. On the next loop, the network inspects state data and conversation history, and determines which new agent to run.<p>This fully typed state machine routing allows you to deterministically build agents using any of the effective agent patterns — which means your code is easy to read, edit, understand, and debug.<p>This also makes handoff incredibly easy: you define when agents should hand off to each other using regular code and state (or by calling an LLM in the router for AI-based routing). This is similar to the OpenAI Agents SDK but easier to manage, plan, and build.<p>Then comes the local development and moving to production capabilities.<p>AgentKit is compatible with Inngest’s tooling, meaning that you can test agents using Inngest’s local DevServer, which provides traces, inputs, outputs, replay, tool, and MCP inputs and outputs, and (soon) a step-over debugger so that you can easily understand and visually see what's happening in the agent loop.<p>In production, you can also optionally combine AgentKit with Inngest for fault-tolerant execution. Each agent’s LLM call is wrapped in a step, and tools can use multiple steps to incorporate things like human-in-the-loop. This gives you native orchestration, observability, and out-of-the-box scale.<p>You will find the documentation as an example of an AgentKit SWE-bench and multiple Coding Agent examples.<p>It’s fully open-source under the Apache 2 license.<p>If you want to get started:<p>- npm: npm i @inngest/agent-kit<p>- GitHub: <a href="https://github.com/inngest/agent-kit" rel="nofollow">https://github.com/inngest/agent-kit</a><p>- Docs: <a href="https://agentkit.inngest.com/overview" rel="nofollow">https://agentkit.inngest.com/overview</a><p>We’re excited to finally launch AgentKit; let us know what you think!

Show HN: I converted my notebook into a searchable database of IT keywords

Show HN: We made a photo search engine for homes for sale

We're a small team of 3 engineers, and wanted to make a better way to look for for properties online. Traditional portals (esp. outside of US) are just a bit rubbish for doing anything other than basic searching.<p>So we've:<p>- Created a crawler that reads through estate agents websites to find homes for sale<p>- Parses that through a series of LLMs and other models to understand each home in depth (e.g. floor type, location, total sqft)<p>- Parses every photo through an embeddings vector space so that people can search for whatever they want.<p>Check it out: <a href="https://jitty.com" rel="nofollow">https://jitty.com</a><p>Currently only for home sales (not rentals), and only in the UK.<p>Some examples:<p>- "Beautiful church conversions up to £1m" <a href="https://jitty.com/for-sale/price-up-to-1000000-gbp/detached-or-semi-detached-or-terraced-or-bungalow-property-type/look-for-beautiful_church_conversions" rel="nofollow">https://jitty.com/for-sale/price-up-to-1000000-gbp/detached-...</a><p>- "Floor to ceiling libraries with a ladder" <a href="https://jitty.com/for-sale/look-for-floor_to_ceiling_libraries_with_a_ladder" rel="nofollow">https://jitty.com/for-sale/look-for-floor_to_ceiling_librari...</a><p>- "Home in london, under £1m, with big beautiful windows" <a href="https://jitty.com/for-sale/london/price-up-to-1000000-gbp/detached-or-semi-detached-or-terraced-or-flat-or-maisonette-or-bungalow-property-type/look-for-big_beautiful_windows" rel="nofollow">https://jitty.com/for-sale/london/price-up-to-1000000-gbp/de...</a><p>- "Bathtubs with an epic view" <a href="https://jitty.com/for-sale/look-for-bathtubs_with_an_epic_view" rel="nofollow">https://jitty.com/for-sale/look-for-bathtubs_with_an_epic_vi...</a>

Show HN: We made a photo search engine for homes for sale

We're a small team of 3 engineers, and wanted to make a better way to look for for properties online. Traditional portals (esp. outside of US) are just a bit rubbish for doing anything other than basic searching.<p>So we've:<p>- Created a crawler that reads through estate agents websites to find homes for sale<p>- Parses that through a series of LLMs and other models to understand each home in depth (e.g. floor type, location, total sqft)<p>- Parses every photo through an embeddings vector space so that people can search for whatever they want.<p>Check it out: <a href="https://jitty.com" rel="nofollow">https://jitty.com</a><p>Currently only for home sales (not rentals), and only in the UK.<p>Some examples:<p>- "Beautiful church conversions up to £1m" <a href="https://jitty.com/for-sale/price-up-to-1000000-gbp/detached-or-semi-detached-or-terraced-or-bungalow-property-type/look-for-beautiful_church_conversions" rel="nofollow">https://jitty.com/for-sale/price-up-to-1000000-gbp/detached-...</a><p>- "Floor to ceiling libraries with a ladder" <a href="https://jitty.com/for-sale/look-for-floor_to_ceiling_libraries_with_a_ladder" rel="nofollow">https://jitty.com/for-sale/look-for-floor_to_ceiling_librari...</a><p>- "Home in london, under £1m, with big beautiful windows" <a href="https://jitty.com/for-sale/london/price-up-to-1000000-gbp/detached-or-semi-detached-or-terraced-or-flat-or-maisonette-or-bungalow-property-type/look-for-big_beautiful_windows" rel="nofollow">https://jitty.com/for-sale/london/price-up-to-1000000-gbp/de...</a><p>- "Bathtubs with an epic view" <a href="https://jitty.com/for-sale/look-for-bathtubs_with_an_epic_view" rel="nofollow">https://jitty.com/for-sale/look-for-bathtubs_with_an_epic_vi...</a>

Show HN: We made a photo search engine for homes for sale

We're a small team of 3 engineers, and wanted to make a better way to look for for properties online. Traditional portals (esp. outside of US) are just a bit rubbish for doing anything other than basic searching.<p>So we've:<p>- Created a crawler that reads through estate agents websites to find homes for sale<p>- Parses that through a series of LLMs and other models to understand each home in depth (e.g. floor type, location, total sqft)<p>- Parses every photo through an embeddings vector space so that people can search for whatever they want.<p>Check it out: <a href="https://jitty.com" rel="nofollow">https://jitty.com</a><p>Currently only for home sales (not rentals), and only in the UK.<p>Some examples:<p>- "Beautiful church conversions up to £1m" <a href="https://jitty.com/for-sale/price-up-to-1000000-gbp/detached-or-semi-detached-or-terraced-or-bungalow-property-type/look-for-beautiful_church_conversions" rel="nofollow">https://jitty.com/for-sale/price-up-to-1000000-gbp/detached-...</a><p>- "Floor to ceiling libraries with a ladder" <a href="https://jitty.com/for-sale/look-for-floor_to_ceiling_libraries_with_a_ladder" rel="nofollow">https://jitty.com/for-sale/look-for-floor_to_ceiling_librari...</a><p>- "Home in london, under £1m, with big beautiful windows" <a href="https://jitty.com/for-sale/london/price-up-to-1000000-gbp/detached-or-semi-detached-or-terraced-or-flat-or-maisonette-or-bungalow-property-type/look-for-big_beautiful_windows" rel="nofollow">https://jitty.com/for-sale/london/price-up-to-1000000-gbp/de...</a><p>- "Bathtubs with an epic view" <a href="https://jitty.com/for-sale/look-for-bathtubs_with_an_epic_view" rel="nofollow">https://jitty.com/for-sale/look-for-bathtubs_with_an_epic_vi...</a>

Show HN: I built a MCP server so Claude can play Minesweeper

Hi! I build an MCP server that allows clients to play Minesweeper. It turns out that Claude is not very good at it (makes obvious mistakes, hasn't won a single game on a 9x9 board after many attempts).<p>I am curious how I can prompt Claude to do better?

Show HN: I built a MCP server so Claude can play Minesweeper

Hi! I build an MCP server that allows clients to play Minesweeper. It turns out that Claude is not very good at it (makes obvious mistakes, hasn't won a single game on a 9x9 board after many attempts).<p>I am curious how I can prompt Claude to do better?

Show HN: I built a MCP server so Claude can play Minesweeper

Hi! I build an MCP server that allows clients to play Minesweeper. It turns out that Claude is not very good at it (makes obvious mistakes, hasn't won a single game on a 9x9 board after many attempts).<p>I am curious how I can prompt Claude to do better?

Show HN: I built a MCP server so Claude can play Minesweeper

Hi! I build an MCP server that allows clients to play Minesweeper. It turns out that Claude is not very good at it (makes obvious mistakes, hasn't won a single game on a 9x9 board after many attempts).<p>I am curious how I can prompt Claude to do better?

Show HN: Codemcp – Claude Code for Claude Pro subscribers – ditch API bills

Hi all! I normally work on the PyTorch project but I've been on baby leave for the past month, so I've been playing around with AI as a user rather than a framework implementor. I really liked the agent experience with Claude Code, but I couldn't really justify spending so many dollars on API costs for random side projects. I already pay for a Claude Pro subscription though, and it turns out you can simulate many of Claude Code's features with an MCP. If you have a Pro subscription, check this out! I think it really captures the Claude Code experience quite well, without forcing you to pay for API tokens.