Malte Landwehr

@eringriffith The article should have led with “both founders were in Forbes 30 under 30” and I would have believed all fraud allegations without question.

A detailed and brutal look at the tactics of buzzy AI compliance startup Delve "Delve built a machine designed to make clients complicit without their knowledge, to manufacture plausible deniability while producing exactly the opposite." substack.com/home/post/p-19…

@JamesIvings @vercel It is a free plan. You are the product. Nothing mad about that.

if you use @vercel free plan, they now train their AI on your code by default 😱 madness

@andyorsow No. Under no circumstances ever.

Does anyone else want a monitor stand that's just a chunk of rock?

Not getting cited by LLMs? You can debug that! Here is the exact framework I use to find the gap - and fix it. Tried and tested across dozens of domains. Let’s assume I have a website and want to be the source for a set of prompts. Each prompt generates some fanout queries. I have already gathered these. With this framework, we will iterate through all prompts and all fanout queries 1️⃣ Is your website directly cited in the answer? Yes? Perfect. Your work is done and you should probably focus on other prompts where you are not cited, yet. No? Go to step 2. 2️⃣ Is your domain used as a source but not explicitly cited? Yes? Make it more citeable / quotable. Add a summary with the relevant key statement. Keep it short. Write in clear and simple language. Keep the entity density high. If relevant, explicitly express relationships between entities. If you use technical terms or abbreviations, shortly explain them. Use an authoritative, declarative voice. The summary does not have to be for the whole article. You can also summarize a long paragraph, a table, or a chart/diagram. It a summary does not fit, you can use question-answer style content blocks. No? Go to step 3. 3️⃣ Does your domain rank for the fan-out query but not get used as a source? Yes? Improve the content so it is more source-worthy. Make sure you agree with the broad consensus. Add unique information no other source has. Strengthen your EEAT signals. Add quotes from trustworthy sources. Explicitly name your sources. Ideally, expand your document to cover multiple fanout queries. No? Go to step 4. 4️⃣ Do you already have a relevant page for the fan-out query that does not rank yet? Yes? Make sure it is indexed. Build internal and external links. Improve the content. Make sure there are no crawling or rendering issues. In short, do SEO. No? Create that page.

@natjin Why would the best AEO/GEO companies focus on Perplexity, which is only the - checks notes - 7th largest LLM-based search and answer engine?

the smartest aeo/geo companies are using perplexity’s search api like a super-weapon they realize you can actually see what is getting ranked and retrieved by LLMs at mass-scale (200m+ daily perplexity queries) vs prompting llms to try to get answers this is real alpha

Most of the magic happens in this table @peec_ai

@GergelyOrosz Which is crazy to me, because Google Flights is by far the best product in the market.

In 2010 people working at Google would have believed Google will kill most startups when they enter a category. Case in point: Google Flights. It launched in 2011. Should have killed most flight comparison websites + travel agents. Yet today they are doing better than ever…

@Kevin_Indig The funny thing is that this is often used by people who were too early. Which means they were wrong. And now that certain things finally matter, they use that to vindicate their wrong advice from 10 years ago.

Good god, do we have an "I said this 10 years ago" complex in SEO! No, I'm not imune. But man...

@afgmantu @michael_timbs 100% deceptive. The whole element only exists to trick people into thinking these are logos of reference customers or partners.

@michael_timbs “People who use our tools also use (toolA, toolB, toolC).” Not deceptive if the language is clear. Similar to “people who watch X also love Y”

Ahahahahahahahahaha

@staysaasy What is the reason? Lack as sales, marketing, and trust?

Re: SaaS death - I actually know of two separate SaaS companies that had employees leave in the last two years to build competitors and in both cases the competitive products are now dead, with zero traction. And the people that left those companies were very, very smart. And the products they built were the same shape as the companies they left, and they used AI to build them. But they had absolutely 0 success.

@metab33d I mean these.

@MalteLandwehr Aren't citations linked inside the answer? I thought you could always see them with a little button in the text

ChatGPT regularly cites deleted Wikipedia articles. And marketers can abuse it! I have repeatedly seen deleted Wikipedia articles in ChatGPT citations. Even 4 months after they have been deleted! The articles do not rank in any search engine (Bing, Google, DDG, Brave). No other LLM (AI Overviews, AI Mode, Perplexity, Gemini, Copilot) cites them. An yet the get cited by ChatGPT over and over again. Below is a chart that show how often an article deleted in November shows up as a source in ChatGPT for a set of prompts I am monitoring with Peec AI. As you can see, the deletion had no impact on how often I encountered the article. This suggests to me that OpenAI has a cache of certain domains - at least Wikipedia. And they do not properly remove documents from it. SEOs have always spammed Wikipedia. I fear this observation will increase such behaviour. Get a Wikipedia article for your brand. Lose it after a few days. Profit from additional AI search visibility for months.

@TechLayoffLover Every bloated company should do the same.

Meta just confirmed a 20 percent company wide headcount cut but insiders are screaming the real story is way uglier than the efficiency memo. Direct word from multiple sources deep in the org. The announced 20 percent is only the near-term floor. A senior manager told me it's the starter number. Top leadership is already modeling two to three times that long term. They dug into developer activity logs after rolling out heavy AI tools across engineering. What they found floored them. Hours worked have plummeted for a huge chunk of the org. Many seniors now clocking under 10 hours of real productive time per week because the tools are handling so much grunt work so fast. One staff engineer said leadership pulled dashboards showing commit rates holding steady or climbing while logged hours tanked. Same output. Way less human sweat. The math was brutal. If AI is letting one person do the work of three with half the time input why pay for three. They have been quietly running knowledge extraction sprints for months. Senior engineers screens recorded 24 7. Every prompt logged. Debugging flows captured in brutal detail. Entire decision trees filmed under the banner of process documentation for continuity. One engineer described being forced to whiteboard his whole system design playbook. Trade offs. Failure modes. Scaling hacks. All while cameras rolled. They called it knowledge transfer to support the transition. The transition is agents and remaining skeleton crew armed with those exact recordings prompt libraries and heavy Claude+Gemini access. Replacements are already shipping changes 40 percent faster using the precise workflows they ripped from the outgoing engineers. They are not slashing for cost alone. They are turning 15 plus years of Meta engineering DNA into structured training data. Every document your process request is not teamwork. It is feeding the beast that replaces you. The strategic AI pivot line they are feeding the press is cover. They are replacing the entire engineering organism with agents trained on their own seniors captured minds. CTO level playbook is already locked. Extract. Document. Automate. Repeat at two to three times scale. If you are still at Meta and someone pings you to please record a quick walkthrough of your workflow for the team run. Do not document. Do not hand over the keys. The knowledge extraction is complete. If you are inside watching more engineers get gutted because AI made their 40 hour week look like 10 hours of value, DMs are wide open.

@ctwtn Will templates and calculators really win? These are two of the use cases I fully moved to LLMs (as a user) an would not even think about looking for on Google.

Most SaaS marketing teams still run the same SEO playbook: Write another “Ultimate Guide.” Rank for “What is X.” Capture traffic. The problem? AI now answers those questions directly. Studies show that when AI summaries appear in search results, click-through rates can drop dramatically for informational queries. So what are high-performing SaaS companies doing instead? They’re shifting from blog-led SEO → product-led SEO. Instead of writing about problems… They’re building solutions that rank. Examples: • Templates • Calculators • Integration pages • Workflow tools • Use-case hubs Why these win: 1️⃣ They solve the job immediately 2️⃣ They’re harder for AI to summarize 3️⃣ They convert far better than blog posts One useful free tool often outperforms three well-written articles. The new SEO question isn’t: “What should we write about?” It’s: “What can we build that our buyers are already searching for?”

@redlewel @beginbot Deployment into production is built in with Lovable. You need zero knowledge about anything and can have a web-app in the hands of your users within minutes after signing up.

@beginbot This times a million even if you are a straight up viber coder who doesn't know programming why wouldn't you just use claude code or codex? No one uses these platforms, except for people getting tricked and wasting $50-$100 in tokens before giving up

Ahh so Lovable uses classic scammer tricks to seem more relevant! This makes sense since I've never met a single person IRL who uses Lovable.

@beginbot Almost every non-engineer starts their vibe coding journey with Lovable. I know dozens of people using it.

@pcshipp .com is best .xzy is worst

Which domain better for SEO? .io .ai .app .xyz .com

@metab33d They are what OpenAI calls sources in the mobile app and citations in the web interface. Not links within the answer I never use the API.

@MalteLandwehr Are these inline citations or marked as sources because they can appear in the training data? This happens frequently when you use the API and request sources.

The only disgusting thing here is the anti-AI and anti-progress framing. Technological advancements have always made some jobs redundant and some skills less valuable. Every single time such progress happened in the past, it was good for society: calculators, cars, automatic phone switchboards, CAD software

@NYMag Absolutely disgusting

“This is my only source of income right now,” a former TV writer says, about his job producing data for AI. “I know people who are award-winning producers and directors, and they’re not advertising that they’re doing this work, but that’s how they’re putting food on the table.” nymag.visitlink.me/3eAN_5

@NYMag Did calculators and cars steal careers? If these writers, producers, and directors are so talented, why don’t they use AI to create something that previously would have taken a huge budget?

@signulll I cannot imagine anything worse than random neighbours wanting to form a community with me I can barely tolerate WhatsApp groups for buildings. And even those only by muting them.

what adam neumann is doing now is such an obviously good idea. most of american life unless scheduled is super isolationist so you rarely get those spontaneous interactions with ppl.. this sorta flips that around right at the point of where ppl spend majority of their lives. anyway, it’s super duper interesting to see the wework model applied to residential.