Webflow’s New OpenAI Codex Integration Simplifies Site Management - Search Engine JournalWebflow’s New OpenAI Codex Integration Simplifies Site Management Search Engine Journal
New Tracer Open-Source App Combines Opus 5 and Codex WorkflowsOrchestrate multiple AI models in a unified workspace using Tracer. The free desktop app lets Opus 5 and Codex collaborate seamlessly on macOS.
Webflow’s New OpenAI Codex Integration Simplifies Site ManagementWebflow adds Codex support and built-in Skills that vastly simplify using AI to manage content, site optimization, and development tasks.
How Much to Charge-Freelance/Contracting – Web/UX/Dev/Marketing Any Niche (2026)How much to charge in 2026 amidst the last few years of world economic chaos/politcal impact on the COL, SOL and market crisis/ AI Implosion/changing client demands?UX - Marketing/Web Design For SMEs but I follow Hacknernews for the Dev guides, the crazy insight from the posts here.Niche, Hourly rate, experinece, tips, Changes during latest world economic issues, etc and anything else (What tools used, native build/versus tools etc)In the end, providing a solution to a pain/problem is the key -but some are zealots for code versus no-code tools. Im not here taking sides- Im just curious about the 2026 update with the AI explosionEntry-medium experience Marketing/UX and No-Code tools ( Tabooooo) - currently learning front end but more focused on UX/Marketing/Solutions fors SMESBased in Oceania (AUS/NZ market) but looking to do overseas clents as Oceania is limited in business opps.
I write my own MarkdownIn the context of side projects, I've come to these rules for code and text:1a) Code gets minimal review so long as I've architected the solution, built around the correct data structures (I've heard the code writes itself if the data structure is correct), and have specified the shape of the tests.1b) Tests I review to check for overfitting, cheats, etc.2) Comments in code can be chock-full of LLM-isms. Don't care.3) But the Markdown files, which may be read by a person or an LLM, those must be heavily redacted, edited, and put into a human voice.The simple reason is that LLM's, machines built to generate text, are incapable of presenting a linear flow of relevant facts without sidebars, tangents, irrelevancies, duplication, gaps, well-known style fails like overuse of stock words and phrases, layers of historical cruft, over-explanations, and qualitative fluff. In other words, they can't tell it straight.I have only one audience, people. Some of them read the text themselves, others
Ask HN: Is there any way to use workflow to control Harness?I have used skills and MCP to configure my Agents, and I heard about many essays about the workflow. But I don't know how to use workflow in agents such as opencode or pi.
Microsoft Fixes Entra ID Flaw that Could Allow Remote Code Execution - Petri IT KnowledgebaseMicrosoft Fixes Entra ID Flaw that Could Allow Remote Code Execution Petri IT Knowledgebase
Princess Kate's strict dress code at the Balmoral dinner tableKate Middleton and other members of the royal family must follow a strict dress code at royal residences such as Balmoral Castle and Sandringham Estate, according to Tanya Rose Simpson ...
Stress-Testing the Options: Low-Code Platforms for Teams Ready to Commit - TechloyStress-Testing the Options: Low-Code Platforms for Teams Ready to Commit Techloy
Sony 1000XM6 Headphones Are Nearly Back to Prime Day Pricing, Now Almost as Cheap as Buying UsedFor noise-cancelling over-ear wireless headphones, that’s the Sony WH-1000XM6, which are now at a $398 Amazon sale price that comes close to their Prime Day record low. No codes or coupons needed, and ...
Zide Opens Public Beta of Its Native AI Development Workspace – Code, Git, Issues, and CI in One AppNative desktop application brings the full development workflow - projects, git, issues, pull requests, CI, terminals, ...
Show HN: linecast – weather, radar, tides and maps in the terminalI've been working on linecast, a free and open source collection of six terminal applications: weather, radar, sunshine, moon, tides, and maps.It's as if the Old Farmer's Almanac was built for Minitel.I made it because I'd been using the terminal a lot more over the past couple of years, I'm opinionated about weather forecast IA and design, and I liked the idea of making useful little terminal desk ornaments.The applications use ANSI color and Unicode braille characters rather than graphics. They reflow as the terminal is resized and derive their colors from your current terminal palette, including when the theme changes while they're running. They're mouse friendly where it makes sense.There's no account or required API key. Forecasts and air quality come from Open-Meteo; alerts, tides, and radar use various public sources; Sun and Moon positions are calculated locally; and the maps use OpenStreetMap-derived vector data and public elevation tiles. The code is just the Python stdlib wi
Gartner: Low-Code Tech is Projected to Grow to Nearly $27 Billion in 2023 - TechRepublicGartner: Low-Code Tech is Projected to Grow to Nearly $27 Billion in 2023 TechRepublic
Show HN: Ditdah – A Morse App for the TerminalI passed the Morse exam when I got my novice license, but I haven't started using Morse until recently. I'm really a beginner: I can key at about 15 WPM and am practicing for 20, but I'm still making mistakes, with an extra dit or dah here and there. Thanks to all the YLs and OMs who have worked with me so far!I'm able to decode Morse with pen and paper, as long as it's only 10–15 WPM, but that needs to improve. Most operators recognize CQ without even knowing Morse. I have already learned a few more phrases just by hearing them, like "TEST", "DE", "PSE K", "BK", "BT", etc. Nevertheless, I have issues with more conversational words like "sunny", but it will build up slowly, I'm sure.To help myself out, and to learn something new, I started building a terminal app for Morse decoding. It's built around a neural network that I trained and can pick up quite a bit of text really well.The interface is textual. These kinds of applications have a special place in my heart, as I grew up in the
Show HN: Oxbow, a dynamic tiling window manager for the Wayland compositor RiverIt's inspired by dwm, but it's configured via an ipc/init script, like river-classic, which makes it a bit easier to get into/test. Specifically, it has:- Tags: Group windows with tag sets, per-tag layouts, and operations on tags.- Layouts: dwm-like tiling, niri-ish scrolling in four orientations, “alt-tab” overview, and pure floating.- Windows: Scratchpads, sticky windows, user labels, window swallowing, and drag with retile.- Rules: Regex window matching to configure initial tags, output, floating, size, and presentation.- IPC: Control windows, configure tags, query objects, and subscribe to events.Maybe you'll use it, or it inspires you to write your own (since River does all the hard stuff) :DRiver: https://codeberg.org/river/riverOther River Window Managers: https://codeberg.org/river/wiki/src/branch/main/pages/wm-lis...
Tell me "Yes" if you are handling any of below AI coding chaos..If you're using Cursor, Claude Code, Codex, Antigravity or any other coding agent..Tell me "Yes" if you are handling any of below AI coding chaos..1. Too many AI agents.2. Copy-pasting context.3. Switching between agents.4. Losing project context and memory.5. Repeating the same prompts.6. Burning tokens on simple tasks.7. Paying for expensive models unnecessarily.8. Not knowing what AI changed.9. AI tools don’t talk to each other.10. Too many tabs. Too much chaos.No worries..We’ll show you what we’re building to fix it. Only 10 seats available. Comment if you are interested.
Show HN: Samuel, a Silly Speech ModelMy original idea was to make a modern supercharged version of SAM, the old-school text-to-speech. Then I remembered Pink Trombone, one of my favorite creative coding projects. In the end, I decided to just make it a model that mimics speech - it would be easy to make a text-to-speech by chaining an existing TTS with a Samuel filter.The phonetics work quite well for vowels - you can try making a sound and then seeing if the predicted tongue prediction matches where your tongue was. It's less accurate for consonants; it doesn't do plosives or nasals (I tried).How the machine learning works:It's a speech autoencoder in which the decoder is Pink Trombone. A couple of issues: PT is not differentiable, so I use Gumbel softmax to get a differentiable approximation. PT also generates audio sample-by-sample (apparently it's called a "Kelly-Lochbaum vocal tract model"), which would make training extremely slow. To do that, I approximate it using FIR filters that change each 10ms. The fundamental
Ask HN: If you dislike AI, why don't you prove it?I've seen so many people who absolutely hate AI, and want AI companies to fail, but then 2 seconds later, open up the coding tool made by frontier labs and willingly hand them money.Just recently convinced one of them to move away from the frontier lab lock-in to OpenCode and switch between models. Not saying they need to switch to non-frontier models immediately, but this approach lets them easily switch to open source when the time is right.Why don't more people do this? And shouldn't companies that use AI heavily also suggest this approach?To be clear, I like AI, and use it all the time, but I use it in a way that allows me to easily switch models and not get locked-in.