If you don't have time between the jobs - how do you switch from a toxic environment to the new one without bringing the negativity and other psychological barriers with you?
We audited what our own site actually answers, and the result came out backwards from what we assumed.
A couple of hundred pages, most of them a template matrix built for search traffic. Those were cached properly. The pages the whole matrix funnels into, the homepage and pricing and the free generators and the legal pages, were answered no-store, every one of them. The pages we care most about were the only ones a shared cache was never allowed to keep.
We fell into the trap of blindly trusting our feature request board. Our top voted item was an advanced reporting dashboard so we dedicated weeks of dev time to make it perfect.
The usage numbers after launch were a reality check. Most people who asked for it barely touched it proving there's a huge gap between what users ask for and what they actually use daily.
I have a rule in my blog that noindexes a category page while it has zero posts in it, and drops it from the sitemap at the same time. I wrote it on 18 July. It is one line: robots index false when the post count is zero.
On 13 August Google crawled one of those category pages while it was still empty. It picked up the noindex, and Search Console filed it as Excluded by noindex tag.
A post landed in that category some time after that. I checked the page today. It returns 200, it serves no robots meta at all, it lists its post, and it is back in the sitemap alongside the other 270 URLs.
Search Console still says Excluded by noindex tag. Last crawl 13 August. Fifteen days of a verdict about a version of the page that no longer exists, and nothing in the report tells you the reason is a snapshot from the last crawl rather than a current read.
Not a launch post. Just things I wish someone had written down before I spent a month figuring them out.
1. LLMs send partial payloads on write operations
You ask the agent to update a record. It sends only the fields you mentioned in the prompt. The PUT request goes through, returns 200, and you've silently wiped every field you didn't specify.
The fix: before every write call, fetch the current resource state via the companion GET endpoint and deep-merge the LLM's payload on top. The LLM only needs to specify what's changing the executor fills in the rest.
My content team used to publish two articles each week. We spent a lot of time researching topics but search traffic barely moved.
Writing content takes weeks to rank and requires building fresh authority from scratch. Meanwhile dozens of posts sat on page two of search results and received zero clicks.
We stopped writing articles for a whole month to focus only on our existing library. We updated statistics added better images and expanded thin sections in past posts.
Search engines already trust URLs so giving them fresh content pushes them up the ranks much quicker. A post that moves from position eleven to position four brings hundreds of visitors overnight.
Tab came out of stealth this week, and TechCrunch reported a $300M valuation. It's a personal assistant you text on iMessage or WhatsApp. The founder said it has its own phone number, computer and wallet, so it does the task instead of telling you how.
I like that there's no app to install. What I'd worry about as a user is checking what it actually did. A chat log isn't a great receipt when money is involved. For anyone building on WhatsApp or SMS, how do you show people what happened after the fact?
Even back in university, I noticed how much younger people (17 or 18 years old) were always jumping into some kind of trendy business model or income stream that happened to be booming at the time.
First, it was dropshipping and flipping items.
Then came NFTs and everything happening in the crypto space.
Now everyone seems obsessed with quick vibecoded AI solutions and investment apartments.
