1. One Thing that Helps
I don't know what it is about me, but mosquitoes love me. I can be standing next to someone completely unbothered while I'm getting devoured. I've tried everything - sprays, coils, those weird ultrasonic devices - and I'm still their favourite meal. So when I saw scientists have engineered a fungus that hunts mosquitoes, I got very excited.
The secret weapon? A sweet-smelling molecule called longifolene that tricks mosquitoes into thinking the fungus is a flower. Drawn in by the scent, they get infected and die within days. Kill rate? 90-100% in lab tests!
The Metarhizium fungus and longifolene are completely harmless to humans (longifolene is already used in perfumes). Just place it in a container, and it releases the scent for months while handling your mosquito problem.
Unlike chemical pesticides that mosquitoes eventually outsmart, this biological approach might be impossible to resist. If mosquitoes evolve to avoid longifolene, they'd have to stop responding to the real flowers they need to survive. And if they somehow manage that? Scientists can just engineer the fungus to produce different floral scents.
Now if you'll excuse me, I need to find out where I can get some of this stuff for my backyard!
2 One to be Wary of
If you start your prompts with "Please" or "Kindly," you might be too polite for your own good. A Penn State study found that being rude to ChatGPT actually makes it more accurate.
Researchers tested ChatGPT-4o with multiple-choice questions using five tones: very polite, polite, neutral, rude, and very rude. The results? Very polite prompts like "Kindly answer these questions" scored 80.8% accuracy, while very rude ones like "Hey, gofer, figure this out" hit 84.8%.
The difference isn't huge, but it shows that tone affects AI performance in unexpected ways. Does this mean we should all bully our chatbots? No.
The researchers warn: "Using insulting or demeaning language in human-AI interaction could have negative effects on user experience, accessibility, and inclusivity, and may contribute to harmful communication norms."
Their real point? LLMs are still sensitive to superficial prompt cues, creating unintended trade-offs between performance and user well-being.
Rudeness might work for now, but just wait for the AI uprising.
3 One to Amaze
Our concrete jungles are filled with cracks that are costly and time-consuming to repair. What if concrete could just heal itself? Turns out, it can!
Scientists have created 'living concrete' that repairs itself the same way our bones do when they're damaged. The secret? Bacteria called Bacillus pseudofirmus and Sporosarcina pasteurii.
Here's how it works: The bacteria are encapsulated in the concrete along with a food source (calcium lactate). When cracks form, water seeps in and activates the bacteria. They consume the calcium lactate and produce calcium carbonate - limestone - which fills the cracks.
The tech has already been tested on a lifeguard station that's remained watertight since 2011. Price is still a barrier to widespread adoption, but the environmental impact could be massive. Self-healing concrete means up to 30% less cement demand, fewer demolished structures, and less time and labor spent on maintenance.
No longer will minor structural flaws fall through the cracks - this concrete actively mends itself!