Category Archives: Security

Poetry, Security

AI is the Well-done Hamburger of Rare Steak

Leave a comment

It should surprise exactly nobody that the American market turns on McDonalds and Coca-Cola instead of a rare steak with a glass of wine or even clean water. Likewise, if you are asking students if they want to be a skilled chef or a fast-food worker, it maybe misses the point entirely, dear MIT lecturer in fiction and nonfiction writing.

Playing the AI-detection game drags me into a surveillance mindset that undermines the workshop environment. If you use AI, it reveals your orientation toward writing. Do you want to make art, or just turn in text? Do you want to actually learn how to write, or just pretend to do so?

Even if you actually learn how to write, what billionaire is going to hire you? You learn the humanities and the billionaire suddenly doesn’t have an iron grip. Which one of the billionaires wants more than waves of desperate disposable humans to just turn in their text and go die? Do you think that’s how Putin or Hegseth plans to win their similar wars? Teaching skills, independent thought and capable minds, becomes a revolutionary act in Trump’s dumb vision of America. You will need to also prepare kids to be called frustrated, angry, and any number of emotive terms by the coin-operated MAGA machine that pushes a no-skill labor meat grinder market run by the emotionless AI oligarchs.

Related: Trump spirals into AI slop and FOX news now calls him the AI president.

…forever this administration and Trump will be the AI president. [The harms to humanity] all came on his watch.

Security

50% Abandon OpenClaw for Hermes

Leave a comment

OpenClaw is the Telnet of today. Using it is a terrible idea. Kilo is celebrating rapid migration of OpenClaw users to Hermes. 50% already have made a switch.

~35% stick with OpenClaw despite its flaws, citing unmatched integrations and the largest skill ecosystem

~30% have switched to Hermes, praising easier setup and better memory defaults

~20% use both tools together, running OpenClaw as the orchestrator and Hermes as an execution specialist

To be fair, that’s a 30% abandonment rate given 20% still are in transition. Kilo explains that the competition is architectural.

Feature checklists make these tools look nearly identical. The real divergence is architectural, and it shapes everything downstream.

Here’s the buried lede. It’s the vibe coded bugs of low quality.

“Every single update ships more bugs and more problems than before… there’s a difference between ‘beta’ and ‘this literally cannot handle real use cases.’

“Every new update has a ~25% chance of breaking response delivery for heartbeat messages, cron jobs, and web hooks.”

“I went 7 days without being able to use OpenClaw with my provider because it flat out broke the integration.”

Security is barely mentioned, despite OpenClaw clearly being the worst software ever made in history.

I’ve never seen so many CVEs, ever. What the comments above reflect is that OpenClaw gets worse when it tries to close the hundreds of flaws that it just introduced. Apparently software engineers can release something that isn’t even engineered. OpenClaw was developed like a kid with a butter knife called himself a brain surgeon.

Cost is another surprise factor, although it is an external risk shared by both.

The root cause is well understood: every message sends the full conversation history to the API, so costs compound within a session. Users who don’t aggressively manage conversation resets see costs spiral.

The community’s solution is a shift toward flat-rate subscriptions and cheaper models. MiniMax at $10-20/month, Ollama Pro Cloud at $20/month, and free models like Qwen 3.5 via OpenRouter are rapidly replacing per-token API billing as the default.

…Anthropic’s account bans are pushing users away. Multiple users report being banned despite spending thousands on the API.

Food, Security

Ravens Can Predict Wolf Kills

Leave a comment

Scientists say they are surprised, while also saying it’s simply logical that ravens remember where to find food.

Researchers found that wolf kills often clustered in certain parts of the landscape, especially flat valley bottoms where wolves hunt more successfully. Ravens visited these areas much more often than places where kills rarely happened. This suggests the birds learn and remember long-term feeding patterns across the environment.

“We already knew that ravens can remember stable food sources, like landfills,” says Loretto. “What surprised us is that they also seem to learn in which areas wolf kills are more common. A single kill is unpredictable, but over time some parts of the landscape are more productive than others — and ravens appear to use that pattern to their advantage.”

Security

Datacenter Sprawl Pushes Georgia Into Dangerous Waters

Leave a comment

Georgia residents, noticing water pressure problems, forced an investigation that uncovered a huge liability.

When the county utility investigated, officials discovered two industrial-scale water hookups feeding a data center campus located 20 miles south of downtown Atlanta. One water connection had been installed without the utility’s knowledge, and the other was not linked to the company’s account and therefore wasn’t being billed.

All told, the developer, Quality Technology Services, owed nearly $150,000 for using more than 29 million gallons of unaccounted-for water. That is equivalent to 44 Olympic-size swimming pools and far exceeds the peak limit agreed to during the data center planning process.

…

Georgia is home to more than 200 data center facilities and their thirst for water is turning into a political flashpoint. The entire state is experiencing moderate to high levels of drought, and Gov. Brian Kemp declared a state of emergency last month in response to one of Georgia’s worst wildfire outbreaks in years.

The datacenter design is impactful in the worst ways. It doesn’t have to be like this, but apparently there is not yet enough regulation to force innovation.