Joachim B. Auster

@westguard Nice reframing for all the real struggle and pain with AI 😉🙌👍

AI provides incredible development tools, but its biggest advantage comes when you know enough to challenge it. The best developers don't just prompt AI. They verify it.

Paraguay 🇵🇾 und Argentinien 🇦🇷 haben bei der ⚽️ WM 2026 mehrfach auf „brutale“ Weise gezeigt, dass sie Respekt und Fairness gezielt mit Füßen treten. Diese beiden Teams haben im Fussball KEINEN Platz. ❌ Final Decission: Red Card 🟥 #FIFAWorldCup #respect #fairness #redcard

@duplicatorwp that’s why I use @duplicatorwp for more than 5 years 🫶👏

👏👏

@srikat @thesamuelnam got the info yesterday in germany 🇩🇪 lines to australia 🇦🇺 are quite long, might take some days 🤣🤣🤣

Anyone in/from Australia have username option in WhatsApp available yet? cc @thesamuelnam

@rilwis @natmiletic 3+ million installations of what? More context always helps.

@natmiletic I can’t say 3+ million installs is underrated.

Duplicate Page is one of the most underrated plugins for WordPress. It allows you to duplicate a page, post, and just about anything else in the dashboard.

@SMExaminer „We are now running a split test on the new design and so far it is winning.“ Any concrete details on this or only vague words without any value?

I was days away from vacation. My product site looked like it was built five years ago. And I had zero budget to fix it. That's when I decided to try something I'd never done before. In the past, I hired a user interface designer. He built the design system. Then I hired a WordPress developer to make it come to life. This process would take months and tens of thousands of dollars. This time, I didn't have the luxury of going that route. I needed something quick and it needed to look good. Then it clicked. There was a new way to solve this problem. A guest on my AI Explored podcast was raving about the relatively new Claude Design. It's capable of creating slides, wireframes, animations and… landing pages. It was a Saturday and I was just days away from going on vacation. No better time to try something new, right? What I experienced absolutely blew my mind. Here's what I did. I downloaded the product logo and masthead image from my AI Business Society sales page. The main image was an illustration of people in the wilderness pointing to a mountain. And the logo looked like a badge with earth-tone colors. I then went to Claude, clicked on the Design icon, and chose Design System. All I did was upload those two images and the URL of the landing page. Before my eyes, I saw the art elements come alive. Claude Design created motifs and surfaces (paper grain, trail rules, brass hairlines, dark fields), core brand colors (barn red, forest green, brass, sunrise ember…), fonts and spacing, buttons and icons, and a whole lot more. It also was able to clearly identify my style of writing. It keyed in on things like sentence rhythm, social proof use, and even outcomes over features. All of this was simply discerned from what little I gave it. What happened next was the big unlock. There was an option to create a landing page. So I clicked on it. It asked me if I would like to see multiple hero sections to select from. Of course I said yes. Then it started building out the full landing page. It wrote all the copy and it was good. It crafted all the sections. It was amazing. I was flabbergasted. I then downloaded the HTML file and sent it off to two of my staffers and texted them. They loved it. I would go on to art direct Claude Design by simply telling it what to change, just like I was talking to a person. We are now running a split test on the new design and so far it is winning. The initial design took less than an hour. Let me know if you want to see the before and after. Have you tried out Claude Design? I'd love to hear from you. - @Mike_Stelzner

@NotionHQ AI and the huge narrative of „you can“ and „endless possibilities“ AI does not deliver persistent results. The whole AI game changes with each new generation and version of AI models. Currently most experts switch between Codex and Cowork depending what the strength is.

AI is moving fast. Most companies aren't. The loudest stories come from the edges - autonomous agents, massive job losses, whole industries about to flip. Read enough of these stories and it feels like you’re behind and everyone else is way ahead. This didn’t feel quite right to us. So we surveyed 6,000 professionals across 10 markets to map where AI at work actually is. Here's what we found 🧵

@MarkJSzymanski There are so many more questions behind your basic approach. Privacy, Backup, Single Source of Truth. So your code solution might work for you. But what is your approach for teams/sharing, everybody has its own folders etc?

Literally built an AIOS 7 days ago and have not left Claude Code since. Do yourself a favor: 1. Organize all of your context into folders. 2. Create API/MCP connectors to all of the services you cannot rebuild yourself. 3. Orchestrate all of your work. This is the future.

@neuralkian @superwhisper @karpathy @WisprFlow hahaha, you are not aware of data privacy problems and risks with Whisperflow. People only say they use it because of affiliate commissions they get from WhisperFlow.

@superwhisper @karpathy switched to @WisprFlow , it's much better.

LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

@NotionHQ you can, it can, he can, she can. But do each of them really do their job? Nobody knows, because AI agents are a huge narrative that make people buy although nobody cares about their results in real business. A lot of hot air 🟥🚨

Meet Austin, Notion’s Workplace Experience Coordinator. He helps keep the SF office running smoothly and makes sure every new hire feels welcome. One painful part of his job has always been seat assignment: shuffling Notinos and Newtinos, one by one, going back and forth with managers, and starting over every two-week onboarding cycle. So Austin automated the entire seating shuffle end to end with custom agents and a Worker. It’s built using Notion’s developer platform, and connects OfficeSpace and Slack to Notion. Austin put it best: “It’s honestly super easy to set up a Notion worker… you’re just talking back and forth in plain English with Claude Code.” He split the job into two agents: Agent 1 watches the “New Hire” database. When a new name shows up, it posts to Slack and asks managers where to seat them. Agent 2 reads the replies and calls OfficeSpace to assign the seats. Now, instead of spinning up a new Slack channel every cycle and handling requests one at a time, Austin sends one message in Slack, and the agents do the rest. Austin: “I’m no engineer… but this saves me hours every time we onboard. It definitely made my life a lot better.” The best agents aren’t fancy, they just give you time back.

@wpslimseo @WCEurope @wpslimseo look, what @SureRankSEO is doing and evolving related to stability, maintainability and reliability. Hope for yourself to get closer, otherwise you might loose the race. 🤷‍♂️🙈

🔥 We're celebrating @WCEurope - Get 30% OFF Slim SEO Pro until Jun 10. #WCEU wpslimseo.com/products/slim-…

@wp_seopress @wp_seopress look, what @SureRankSEO is doing and evolving related to stability, maintainability and reliability. Hope for yourself to get closer, otherwise you might loose the race. 🤷‍♂️🙈

5★ SEOPress review: Solid work, I love this SEO plugin. — This is at least as good as Yoast or All in One, if not better, and it's a great value. Get the Pro license, it's worth it! wordpress.org/support/topic/…

@NotionHQ Scheduled Agents are the new cashcow for Notion. So you publish one shiny story after another without real numbers of success and only with unproofed claims. Only to feed your cashcow. And businesses run into problems of security, maintainability and consistency. 🟥🛑

Meet Grug. Grug is a “Caveman Designer” Custom Agent that our team turns to for blunt, anti-complexity product/design feedback. Basically: “complexity bad, make simple, make button look like button.” Here’s what Grug said about a recent All Hands.

@NotionHQ @samhli_ Scheduled Agents are the new cashcow for Notion. So you publish one shiny story after another without real numbers of success and only with unproofed claims. Only to feed your cashcow. And businesses run into problems of security, maintainability and consistency. 🟥🛑

“I’ll send notes after the call.” (They didn’t.) So we built an agent that does it automatically. Before a call, it pulls together the docs, decision log, and last week’s notes. After the meeting, it writes the recap and drops it in Slack. Pull up a chair. @samhli_ walks you through the full build.

@GijoVarghese_ This happens when @wordpressdotcom Tech Nerds think they are designers and make such design decisions 😂😂😂🟥❌🛑

WordPress 7 definitely looks much more modern now 😄 I'm not a UI expert, but something about the UI still feels a little weird to me. The background colors and contrast of notices, buttons, links, etc. feel a bit off.

@alexjpanagis I would never trust third party software like you do. Especially for compliance, security and functionality. It could look nice on the top. I will never update live production sites with major updates the same day, because of too many dependencies etc. Never ever. 🟥🤷‍♂️

Updated to WordPress 7.0 across all our sites... And, as usual, everything went smoothly 🎉

@wpfusion will wait for 7.02 or 7.03 before even checking the new major version 7.0

WordPress 7.0 = major release, not a Tuesday patch. Before production: → backup → check plugin compatibility → staging → test checkout / signup / CRM sync Our guide for site owners + agencies: wpfusion.com/tutorials/word…

@WPTutz 100% 💯👍

Possibly an unpopular opinion - but, if you're a developer, it does not matter how good your product is if the educational content behind it is poor on non-existent - you lose potential customers fast. Docs are good - video guides and quick starts rule!

@superwhisper @superwhisper would be nice and helpful if you send emails about new videos on YouTube and new releases 🫶🙌

Superwhisper for Windows just got even better. Version 1.3.9 is live ✨

@NotionHQ copy paste plan mode from claude to notion. Claude went worst in the last 1-2 weeks, latest with Opus 4.7. When will notion follow? 🙈💣

Introducing Plan Mode. Because a prompt without a plan is just a wish. In plan mode, your agent shows its work in a read-only sandbox before touching anything. Just plan → approve → execute.