Rethinking how educational videos are discovered.
Building a credibility-first short-form learning platform where behavioral evidence matters more than pure engagement.
🌐 https://t.co/mcCBkgv9x9
📚 https://t.co/XLW1l3yK9S
Next step: scientifically verifying TechShortsApp.
Moving from a conceptual model to empirical validation—testing how much evidence behavioral signals actually provide about video credibility against expert-rated ground truth.
@techshortsapp
New article! 🥳
How Much Credibility Should Each Signal Contribute?
How TechShortsApp weighs behavioral signals without mistaking engagement for truth.
https://t.co/ox2GaP6kSb
First article published! 🥳
Engagement Is Not Evidence: Rethinking Video Credibility Signals
It explores why engagement metrics don’t equal helpfulness or credibility—and what better signals TechShortsApp could use instead.
https://t.co/ofUSCD5lNS
Shipped!
Techshortsapp v2.4.0 removes design friction and responds better across screen sizes. The core question remains:
Can a platform curate videos by credibility, not visibility?
Next exploration:
Are behavioral signals context-dependent?
Building Systems Under Uncertainty
Every paper I read changes the question.
I started with:
What behavioral signals should a credibility system use?
Now I’m asking:
What cognitive processes produced those behavioral signals?
It’s reminds me of finding more upgraded questions in research than answers.
Spent the last week refining my TEDx talk and had the interview yesterday.
The biggest takeaway wasn’t the interview itself.
It was the opportunity for networking because builders are not never made alone.
Bayes' theorem is used to update the probability for a hypothesis as more evidence or information becomes available. Bayes' approach to probability was notably different from the frequentist approach, which dominated statistical methods at the time.
Documentation is important for any research/product.
Almost finished writing 17 pages, 3 left, and will publish the
documentation in a short time.
It will explain its core belief with the active current version and the archived version.
— TechShortsApp Team
Most feeds optimize attention.
We’re training a system to optimize credibility.
V2.0.0 is invite-only while the intelligence matures.
If anyone is interested trying early, you can request an access.
Thank you.
V1 ranked videos by likes & views.
V2 ranks videos by credibility.
Research ✅
Coding ✅
Now entering testing & observation.
We don’t rank truth.
We measure confidence and update it with evidence.
@techshortsapp
We just shipped one of the most requested features 🥳,
you can now Like and Unlike posts on TechShortsApp!
This is the first step towards building deeper engagement
and giving builders instant feedback.
Visit: https://t.co/FXGBz9YYQm