Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde) header
Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde) profile picture

Dr. Web. seit 1997. Die beste Agentur & Software für Ihr Unternehmen finden. Fortbilden. Informiert entscheid

Freiburg im Breisgau
Joined January 2009
Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)Photo by Dr. Web - Das Fachportal für Entscheider im Digita (@drwebde)
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
Ralph may be the most important thing to learn in AI right now. Someone used this technique to deliver an entire contract that would have cost $50,000 to outsource. Total API costs: $297. At a Y Combinator hackathon, teams shipped 6 working repositories overnight using the same approach. Geoffrey Huntley, who invented this, built an entire programming language over 3 months while barely touching his keyboard. Ralph is a bash loop that runs an AI coding agent repeatedly until a task is done. That’s it. While you sleep, while you eat dinner, while you do anything else, the loop keeps going. It picks up a task, builds it, checks if it works, saves the progress, picks the next task. When you wake up, features are finished. Why this works when normal AI coding fails: most people open Cursor or Claude with a vague idea and no structure. 45 minutes later they’re fixing the same bug for the third time. The AI forgot what they were building. The context got polluted. They’re frustrated and nothing shipped. The problem is task size. One feature has 20 pieces. The AI tries to hold all of them at once. It can’t. It hallucinates. It contradicts itself. You end up babysitting. Ralph fixes this by breaking work into pieces small enough that the AI finishes each one before it forgets what it’s doing. Each loop iteration starts fresh. Clean context. No accumulated confusion. Memory persists through git commits, a progress file, and a task list. The AI reads what happened last time, learns from mistakes, picks the next task. The workflow: Step 1: Describe what you want in plain language. Talk for 2-3 minutes. “I want users to filter tasks by priority. High, medium, low. A dropdown with all three options plus ‘all’. Selecting one filters the list immediately.” Then tell the AI to convert your rambling into a formal requirements list. Step 2: Break requirements into atomic tasks with binary success criteria. Good: “Add a priority column that defaults to medium.” Bad: “Make it work well.” The AI needs to know when it’s done without asking you. Pass or fail. Yes or no. Step 3: Run the loop. Ralph grabs a task, builds it, runs tests, commits if it passes, moves to the next one. You set a limit (10 rounds, 14 rounds, 20 rounds). It runs until everything passes or hits your limit. The math is brutal. A typical Ralph run: 10 iterations, roughly $30 in API costs. A senior developer costs $400-600 per day fully loaded. If Ralph gets you 90% there and you spend an hour on cleanup, you just converted an 8-hour workday into 1 hour plus $30. That’s $500+ of labor arbitrage per feature. Compound that across a month of building. Ralph can replace the majority of outsourcing for greenfield projects. The technique is “deterministically bad in an undeterministic world,” meaning when it fails, it fails predictably. You tune it like a guitar. When Ralph screws up a specific way, you add a guardrail to the prompt. The failures become instructions. The real shift is in role. You stop being the person who writes code. You become the person who writes requirements, defines acceptance criteria, and reviews output. Product designer, not engineer. “Software development as a profession is effectively dead. Software engineering is more alive than ever.” The people who understand systems, architecture, requirements, and quality will thrive. The people whose value was typing code faster are already obsolete. The gap forming right now: builders who know Ralph are shipping 5-10x more than everyone around them. Nobody understands how. The technique is free and open source. The barrier is just knowing it exists and spending 30 minutes to understand the pattern. Three months from now this will be in every YouTube tutorial and paid course. The window is right now, while most developers are still prompting Claude one function at a time and wondering why AI “doesn’t work for them.” github.com/snark-tank/ralph
78
237
26
2,702
610,859
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
🚨🇨🇳 BREAKING: China is generating electricity in the sky. The S2000, the world's first megawatt-scale airborne wind power system, rose 2000 meters above Yibin, Sichuan, generating electricity and connecting to the power grid for the first time.
362
1,758
205
10,881
1,292,833
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
Google's AI Overviews = theft 🚨 And Google just got smacked with a lawsuit by Chegg -- and it's utterly fantastic ⚖️ I'm a former commercial litigator turned online travel publisher -- here are my notes from reading the complaint (link in next tweet) 👇 (warning, this summary is long & took me all morning to put together -- if you're a publisher please take 10 minutes to actually read it & share it with others) 1) Chegg is represented by the Susman Godfrey firm -- which means they are not messing around. Susman has quite the reputation as savvy, hard-charging litigators 2) Importantly, the compliant is grounded in antitrust law (NOT copyright, as I initially expected). And, goodness, it's a real banger .... 3) The complaint says the "exchange of access for traffic is the fundamental bargain that has long supported the production of content for the open commercial Web." 🎯 4) It continues: "But in recent years, Google has begun to tie its participation in this bargain to another transaction to which Chegg and other publishers do not willingly consent. As a condition of indexing publisher content for search, Google now requires publishers to also supply that content for other uses that cannibalize or preempt search referrals." 5) Chegg talks not just about AI Overviews, but also references featured snippets 6) 🫶this part -- "Google’s foray into digital publishing is designed to make Google a destination, rather than a search origination point to other websites." 7) Even if Google provided a separate opt-out for AI Overviews, it wouldn't work - because Google's monopoly power creates a collective action problem 8) A lot of fantastic points about how Google unfairly leverages its monopoly power to bend the open web to Google's will 9) Paragraph 13 is chef's kiss 🫰... "Google’s conduct is already eroding incentives for Chegg and other publishers to produce such valuable and useful content. If not abated, this trajectory threatens to leave the public with an increasingly unrecognizable Internet experience, in which users never leave Google’s walled garden and receive only synthetic, error-ridden answers in response to their queries—a once robust but now hollowed-out information ecosystem of little use and unworthy of trust." 10) Search engines are supposed to be intermediaries between users and web publishers. The complaint quotes old-school Google saying "We may be the only people in the world who can say our goal is to have people leave our website as quickly as possible." 11) Search is a uniquely important channel for publishers. Other channels of traffic like social media can never replace search, because that traffic is not intentional 12) "Put simply, Google’s search monopoly gives it control over online distribution for digital publishers. Google uses that power to force digital publishers to give up their content. Google then itself acts as a publisher, either by republishing portions of other digital publishers’ content or by using GAI to summarize the content. The end result is that users increasingly consume other web publishers’ content on Google’s SERP, either in abridged or derivative form, which starves those publishers of traffic and revenue." 13) Chegg calls Google's strategy for publishers "embrace, absorb, extinguish" 14) Next is an entire fantastic section recounting the history of "Google’s Transformation from a Search Engine to Web Publisher" 15) Google appropriated publisher's content in 2 phases. Phase I it calls the 'republishing phase.' The complaint talks about featured like featured snippets & People Also Ask 16) Hilariously, paragraph 63 includes a screenshot of Sundar Pichai's mug inside a SERP for "who is google's ceo" 16) "Google refers to Featured Snippets, Top Stories, and People Also Ask as “search features.” But they are separate and distinct products from search results. This is Google acting as an answer engine—not a search engine." 17) Republishing is not automatically bad, but the problem is that Google forces it on publishers as a condition for appearing in Search (of which it has a monopoly) 18) "The decision to opt out of republishing by disallowing snippets or withholding Search Index Data is a Hobson’s choice. " 19) Phase II of Google's strategy to dominate online publishing centers around AI Overviews ("GAI") 20) Complaint notes that the non-monopolists like OpenAI and Perplexity have been forced to do licensing deals with many publishers, whereas Google has largely avoided this cost thanks to its search monopoly 21) Complaint walks through the differences between LLM pre-training and RAG (a point I've been trying to educate publishers on for months and months) 22) Next is a recap of Google's LLM product history: Bard, SGE, Gemini & AI Overviews 23) @rustybrick commentary is cited quite a bit (see footnotes on pages 40, 41) 24) The transition to AI Overviews "all but completes Google’s evolution from a “search engine” to an “answer engine” that publishes answers to user’s queries. Its formerly symbiotic and complementary relationship with publishers has now become overwhelmingly parasitic and competitive." 25) Google's own marketing language directly admits that the purpose of featured snippets & AI Overviews is to prevent users from clicking throughs to publishers 26) Next section talks about how Google's unauthorized use of publisher content for AI training 27) "Google has been intentionally vague in identifying the precise data sets used to train the LLMs underlying Gemini and AI Overviews" 28) "Google’s Terms of Service indicate that it uses all the information that it collects for search indexing to train its LLMs, including Chegg’s data. " 29) Google announced Google-Extended in Sept 2023, but blocking it doesn't change that Google trained on your content or that Google uses it for RAG 30) Next section: Google poses a "fundamental threat" to online publishing 👏👏 31) Lots of discussion on how AI Overviews directly compete with and seek to replace online publishers 32) Oh WOW - in paragraph 121, Sussman dug up a pretty damning admission in a 2023 Google DeepMind presentation: Google admits Generative AI in search would "reduce referrals to content providers hurting their ability to monetize" 33) Lots of outside observers recognize the risk AI Overviews present to publishers. Footnotes cite @timsoulo among others 34) AI Overviews increase "zero click" searches 35) AI Overviews will cause a downward spiral in publishing quality - as the incentive to create gets less, quality degrades, and the whole web suffers 36) Google has put publishers in an awful position - by publishing content, they feed the very AI beast that is consuming publishers alive 27) Next, the complaint walks through the actual legal claims -- which, IMHO, are quite strong 28) First - "reciprocal dealing," an antitrust concept which "occurs when a firm with market power refuses to sell product X to a customer unless that customer agrees to sell (or give) product Y to it. In this case, the product Google is selling to (and threatening to withhold from) digital publishers is Search Referral Traffic." 29) Second - "monopoly maintenance." Basically argues that Google's use of AI constitutes a form of "rent extraction" on publishers. 30) Google is using AI to illegally entrench its search monopoly 31) Third - "Unjust enrichment" - which basically means Google has unfairly benefited at the expense of publishers via wrongful conduct. 32) "The value of Google’s models and AI products is directly related to the quality of the works that it acquires to train them and ground their outputs." 33) Finally - a recital of the counts: I - Reciprocal Dealing in Violation of Section 1 of the Sherman Act II - Reciprocal Dealing in Violation of Section 2 of the Sherman Act III - Tortious Conduct in Violation of Section 2 of the Sherman Act IV - Unlawful Monopoly Leveraging in Violation of Section 2 of the Sherman Act V - Unlawful Monopolization in Violation of Section 2 of the Sherman Act VI - Unlawful Attempted Monopolization in Violation of Section 2 of the Sherman Act VII - Common Law Unjust Enrichment 34) Lastly, Chegg's request for relief is of course damages and attorney fees, but also for a permanent injunction preventing Google from engaging in unlawful conduct 35) Chegg demands a jury trial ***link to complaint in next tweet*** My overall takeaway? This is a truly fantastic complaint. The only thing I think it missed is Google's admission in a blog post in summer 2023 that robots(.)txt is not a sufficient consent mechanism in the AI age. But, other than that, they really did their homework and hammered a ton of points I've been railing about for years. If you are a publisher - please share so others see! 👏👏👏
52
238
79
741
98,429
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
CSS position: sticky, scroll-snap + progressively enhanced scroll animation 📜 li { animation: bright; animation-range: calc(50% - 1lh) calc(50% + 1lh); animation-timeline: view(); scroll-snap-align: center; } @​keyframes bright { 50% { opacity: 1; }}
37
263
12
3,463
367,058
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
Wie lange hat es im römischen Reich gedauert, um von London nach Rom zu kommen und wie viel denarii hat es gekostet? 30 Tage, 3000km u.a. mit dem Esel Ein Routenplaner für das Römische Reich Oh mein Gott, diese Website ist ein fantastisches Spielzeug! 😍 orbis.stanford.edu/
100
478
44
3,072
253,413
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
CSS only animated exclusive accordion update 🚀 details { interpolate-size: allow-keywords; } ::details-content { transition: content-visibility, height; transition-behavior: allow-discrete; } [open]::details-content { height: auto; } <details name=accordion>
Animated exclusive accordions with CSS 😎 ::details-content { transition: content-visibility, height; transition-behavior: allow-discrete; } [open]::details-content { height: calc-size(auto); } <details name=accordion> <summary>Exclusive</summary> </details> On the way! 🫶
10
79
740
51,205
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
Stop designing ‘Are you sure?’ modals
139
669
70
12,091
815,106
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
💎 6 Types of Login Screen Six different ways of designing a login screen, each with its strengths and weaknesses.
4
53
1
517
70,888
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
Hervorragende Lektüre für alle SEOs. Rand Fishkin vergleicht die Google-Suche in den USA mit der in der EU, die Auswirkungen der Einführung von KI-Übersichten, Null-Klick-Suchen und vieles mehr.
NEW research: sparktoro.com/blog/2024-zero… - Zero click searches in 2024 - EU and US compared - Searches per searcher over time - Clicks per search over time - Impact of May's AI Overviews on searches & clicks It's all there. Sharing much appreciated 📉🤩😎
2
180
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
Hey everyone! Just launched Tabler Illustrations! 🚀 We spent months working on it, and I'm really proud. 😍 50 SVGs to make your designs pop with auto dark mode and custom colors. Let's take your visuals to the next level! 🎨 tabler.io/illustrations
4
10
52
2,803
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
As a designer, I'm Scared and Excited at the same time Tools You Need to Know About - (Day 97)
9
83
4
1,100
199,468
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
💎 Design System Generator for Figma This is a nice helper for setting up a quick design system in Figma. It helps with variables for color, typography, spacing, grids, radius, and shadows.
4
82
4
611
52,859
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
💎 Creatie: visualize product ideas, fast and easy with AI Creatie enables you to generate UI designs using an AI wizard and edit them in a web-based editor. The tool is particularly useful when you want to explore different ideas quickly.  creatie.ai/
1
27
1
197
17,412
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
💎 Free UI components for Figma & Webflow Flowspark offers a collection of ready-to-use UI components for Figma and Webflow. Hero section, contact page, FAQ, footers, etc. Copy & paste any component to Figma/Webflow with one click. build.flowspark.co/
2
50
1
296
19,157
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
#WordPress Dashboard Dark Mode with a free code snippet and a full tutorial on customising it to get the perfect colour scheme. learnbricksbuilder.com/custo…
7
9
651
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
If you're using Akamai, Amazon Cloudfront, Cloudflare, Fastly, or Shopify, this post is a must-read. Among other things, @cliffcrocker provides clear examples of how you can use server-timing headers to get more visibility into your CDN. speedcurve.com/blog/server-t… #webperf
8
26
1,997
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
Whoa, Google is now showing a separate tab for 'Web Results'. Could this be a look into the future of Search once SGE is ready for prime time? IMO standard web listings will always have their place in search results – interesting they are segmenting in this way.
10
48
8
186
22,400
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
You don't need ChatGPT to build custom apps. Custom app builder - uibakery.io Code accelerator - codeium.com Chat bots - chatsimple.ai Git commands - gitfluence.com Code from sketch - sketch2app.io
46
69
5
290
25,583
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
🔥 The new #WPVulnerability 3.1.0 will notify you if the plugin you are using hasn't been updated in a year, and if it's closed in the WordPress repo. wordpress.org/plugins/wpvuln…
3
9
1
30
1,657
Dr. Web - Das Fachportal für Entscheider im Digita retweeted
All of these work GREAT with Scroll Driven Animation btw 🤓 transition.style
CSS tip: on transition.style by @argyleink ​you can find 46 drop-in CSS transitions ✨
4
68
497
47,732