MORE FROM Christian StandalAfter a long season for the Longhorns, they might have hope with Arch Manning going forward after today's performance. The first WNBA mock draft is out, and it gives an early look at where top players ...
Show HN: PixelPointingBenchmark – Simple tests reveal surprising gapsWe built a small open-source benchmark to test how well vision-enabled LLMs handle pixel-level pointing on screens. Instead of complex UI screenshots, we use synthetic images with basic shapes and clean backgrounds to isolate spatial reasoning and coordinate accuracy.The results were surprising:Many top models miss by tens to hundreds of pixels on trivial tasks (e.g., center of a purple circle or red square). High run-to-run variance in some models (different answers on the same image/prompt). Performance flips dramatically with resolution or aspect ratio changes. Claude Sonnet and Claude Haiku are consistently near-perfect (0–1px error), while others show clear gaps. We wrote a detailed blog post about the findings: https://autodevice.io/blog/wheres-the-pixel-part-1Repo (easy to run, add tests, try new models): https://autodevice.github.io/PixelPointingBenchmark/Curious to see how the latest vision LLMs do on this. If you run it, share your results or feedback.Happy to discuss improve
Outperform Crossword ClueIn case you've faced some hurdles solving the clue, Outperform, we've got the answer for you. Crossword puzzles offer a fantastic opportunity to engage your mind, enjoy leisure time, and test your ...
Julia vs. NumPy performance: Strategy for For-loop?Lately, read about Julia Langs' preference in some quantum research. In general, for single shot computations like matrix and vector operations including linear algebra, the performance between Julia and NumPy is comparable due to NumPy's underlying C/Fortran code base. But for Monte Carlo methods for-loops are hard to avoid. Is there any strategy for For-loops in NumPy avoiding python overhead? Or migrating to Julia is a better choice? I am trying to avoid Rust as I see it as more system level language.
Yarbo's Pop-Up Signals the Future of Smart Snow TechIn Aspen, winter isn’t just a season, it’s a lifestyle. Snow-covered streets, designer storefronts, mountaintop homes, and a global crowd that expects the best of everything, even in the harshest conditions. So when a sleek, autonomous machine began clearing snow on a mountaintop deck, people did what came naturally: they stopped, stared, and pulled out their phones.That machine was the Yarbo Snow Blower, and its winter pop-up in Aspen quickly became one of the most talked-about moments of the season.A Winter Pop-Up That Turned HeadsFrom December 12–18, Yarbo hosted a winter pop-up in Aspen, featuring one indoor pop-up store and a dedicated outdoor demonstration zone. The setting was deliberate. Aspen represents the pinnacle of luxury winter living, a place where innovation, design, and performance are expected to meet real-world demands.Yarbo didn’t rely on static displays or polished videos. Instead, the brand brought its technology directly into the snow, staging a live outdoor demo
Show HN: Wafer – Profile, inspect assembly, and iterate on CUDA within your IDEHi HN, I’m Emilio. We’re launching the Wafer extension for the popular IDEs (VS Code, Cursor and Antigravity).Wafer exists to make performance engineers more efficient. Most of the work perf engs do is extracting signal and turning it into the next experiment. You spend hours per kernel doing interpretation and bookkeeping: which counters matter, what changed, what hypothesis you’re testing, what to try next.Wafer is building an environment where profiling, compiler analysis, and docs are first-class context in your workflow, so iteration is cheap. long-term, that same structured context becomes the interface for an automation layer that can read the evidence, propose a change, and rerun the loop.NVIDIA has poured an insane amount of truth into their tooling. NCU, compiler output, SASS, the counters, the sections, the warnings, the “this is why you’re slow” breadcrumbs. Serious perf engineers already live in this stuff. The real problem is that it’s still not packaged as a tight loop.
Why Every Google Ads Account Needs To Run ScriptsGoogle Ads scripts quietly prevent overspend, tracking failures, and performance leaks that automation and Smart Bidding won’t catch on their own. Most PPC marketers love talking about automation, ...
Show HN: TethysRT, a tiny dynamic loader for 32 bit ARM embedded systemsI’ve been working on a lightweight runtime loader for Cortex-M (and 32-bit ARM more generally) that can load and relocate ELF modules from SD/ flash/ generic storage at runtime. It’s not a dynamic linker in the Linux sense, but it does perform symbol resolution and relocation against a host-defined ABI.The goal is to enable extensible firmware platforms without reflashing or statically linking everything into the main image. I’m sharing it both in case it’s useful to others, and to get constructive feedback.
Vertex Minerals lifts Reward Gold Mine output as underground stoping and plant performance improveVertex Minerals Ltd. ( (AU:VTX)) has issued an announcement. Vertex Minerals reports a steady ramp-up in gold production at the Reward Gold Mine as underground stoping ore volumes and grades improve, ...
Built a content system that 6x'd traffic. Turning it into product. Want to test?Hey YCJumping in here because we've been building something we think is pretty cool and are looking for some founders to test it out and give us some honest feedback.We're technical founders. We hated content marketing. But we needed organic growth, so we built a cohesive system to handle it for us.Researched trending keywords & competitors. Proactively queued topics. Drafted SEO & LLM optimized content. Published directly to CMS. Tracked what ranked. Doubled down on what worked.It's led to 6,000% traffic growth in 6 months for us.Now we're packaging that entire workflow into Averi — a content engine for founders who'd rather be focused on shipping product than writing blogs.What it does: → Researches and queues topics for you → Drafts content optimized for Google + LLM citations → Publishes to your CMS → Tracks rankings and performance → Recommends what to create nextYou approve (and edit collaboratively with your team if you'd like). It runs. Visibility compounds.We're looking for ~5
The offline geocoder we wantedWhat is this?This is an offline reverse geocoder written in Python. Given a latitude–longitude pair, it returns the correct administrative region such as country, state, or district without calling any external APIs. This avoids API costs, rate limits, and network dependency.Why build another reverse geocoder?Most offline reverse geocoders rely on nearest-neighbor lookups. While fast, this approach often fails near borders because the closest location is not always the correct administrative region. This project focuses on correctness over proximity by verifying which boundary a coordinate actually falls inside.How does it work?A KD-Tree is used to quickly shortlist nearby administrative boundaries. For those candidates, the system performs polygon containment checks to confirm the true region. It supports both single-process execution for small workloads and multiprocessing for large batch processing.PerformanceThe system processes 10,000 coordinates in under 2 seconds, with an averag
Choosing the Right Stainless Steel Utility Sink for Heavy-Duty UseIn industrial and commercial environments, every piece of equipment plays a role in keeping operations efficient, safe, and hygienic. Among these essentials, the utility sink often works quietly in the background, yet it supports countless daily tasks. From manufacturing plants to commercial kitchens and maintenance workshops, choosing the right stainless steel utility sink for heavy-duty use can make a measurable difference in workflow and durability.Heavy-duty environments place higher demands on sinks than residential spaces. They are used for washing tools, handling raw materials, cleaning components, and managing waste liquids. A well-designed sink must withstand frequent impact, exposure to chemicals, and constant moisture without compromising performance or hygiene. This is where stainless steel stands out as a preferred material across the metal fabrication and industrial equipment sectors.One of the first factors to consider is material grade and thickness. Industrial-grade st
Show HN: I built an AI portfolio manager and entrusted it with $50kMy investing performance so far was, quite honestly, mediocre. It’s not like I’m making terrible investment decisions, it’s more about not being able to consistently manage my portfolio. I’m getting distracted by other projects, leaving my portfolio in free float, which, ultimately, doesn’t work great.So, the idea was to create a reasonable portfolio manager that watches your portfolio, and alerts you (relatively infrequently) when the action is needed. I’m not looking to YOLO to the moon, I’m looking for some reasonable, middle-of-the-road advice and consistent monitoring. This is why I built my portfolio manager, https://portfoliogenius.aiIn the spirit of putting my money where my mouth is, I have $50k to manage since July this year. Currently, it’s about 7.5% up, which is consistent with the goal I gave it - moderate risk portfolio, with some crypto exposure. Yes, S&P was up 12%, but it is expected given that I aimed for lower risk, and a large portion of the portfolio is bonds and
Show HN: Sub-microsecond (890 ns) trading execution research systemI am sharing a research-grade, open-source trading execution framework that achieves a median end-to-end decision latency of 890 nanoseconds on commodity hardware.The project is designed for education, systems research, and latency instrumentation, not for live trading. It focuses on understanding exactly where every nanosecond goes in a trading execution path.Key features:- Kernel-bypass networking: Direct userspace access to NICs via custom drivers, 20-50 ns RX latency - Lock-free SPSC/MPSC queues: Zero-copy architecture - SIMD feature extraction: About 40 ns per update using AVX-512 - Deterministic replay: Bit-identical execution paths, SHA-256 verified - Nanosecond-level metrics: Full audit logs and performance dashboardTechnical stack: C++17 and Rust, NUMA-aware memory allocation, cache-line alignment, inline assembly for hot paths.The framework is modular, allowing experimentation with different NIC drivers, feature extraction pipelines, or order-flow models such as Hawkes proces
Should You Floss in the Morning or at Night? Dentist Reveals Which Is BestIt's not hard to perform the basics of a healthy dental routine, but patients still have a hard time with one step in particular. When it comes to flossing, people often just aren't doing it — or they ...
Show HN: Titan – JavaScript-first framework that compiles into a Rust serverHi HN,I built Titan, a backend framework where you write routes and logic in JavaScript, and the CLI compiles everything into a single Rust + Axum binary using the Boa JS engine. No Node.js is required in production.The idea is to keep JS developer experience while getting Rust performance and a self-contained deployable server.Current features:JS route DSLAction system mapped to Rustesbuild bundlingGenerated Rust server with AxumHot-reload dev serverSingle-binary outputRepo: https://github.com/ezet-galaxy/-ezetgalaxy-titanWould love feedback on the architecture, DX, and whether this hybrid JS→Rust approach is useful.Thanks for reading!
Show HN: We open-sourced our internal tool for scoring PRs with Claude AIHi HN! I'm sharing MergeMint, an open-source platform we built at TextCortex AI to automatically evaluate and score pull requests using Claude AI.The problem we had: Our engineering team lacked visibility into contribution impact, and performance reviews felt subjective. We wanted objective metrics without the overhead of manual PR reviews.How it works: When a PR merges, MergeMint analyzes the diff, linked issues, and commits using Claude AI. It classifies component/severity, calculates a score, posts feedback on the PR, and updates team leaderboards.The results after 6 months: • 3x increase in developer engagement with our internal bug bounty program • Team members actively compete on leaderboards (gamification really works!) • Performance reviews now backed by concrete contribution data • We caught and rewarded high-impact work that was flying under the radarThe scoring is fully configurable – we defined our components (auth, API, UI) with multipliers and severity levels (P0-P3) with