The new AI Camera Assistant* with Xperia Intelligence brings stories to life. Using subject, scene and weather, it suggests expressive options with adjustments of colour, exposure, bokeh, and lens for breathtaking photos*.
https://t.co/zgSQ9MLWFP
#SonyXperia#Xperia1VIII
Sure-fire way:
- Sleep early.
- Wake up early in the last 3rd of the night - and beg Allah to aid you.
- Go back to sleep.
- Wake up for Fajr.
- Recite Qur’an.
- Drink a cup of Espresso.
- Start the work and don’t have any meal until Dhuhr.
Do this consistently for one full week (7 days).
The flow state you’d be in would be crazy
feynman’s “one week” line wasn’t anti-math, it was anti-bullshit.
he loved math — just not the kind that floats so far from reality that nothing in the universe can actually behave that way.
physics has a built-in filter: if nature disagrees, you’re wrong. math doesn’t have that problem.
github just admitted flat-rate ai subscriptions are dead.
they paused new sign-ups for Copilot Pro. imposed strict session limits and weekly usage caps.
the reason?
agentic workflows generate backend costs that exceed a single user's monthly plan price.
a small cluster of parallel requests can burn more compute than github charges for the whole month.
the new limits:
Pro tier gets tight boundaries.
Pro+ gets 5x the capacity.
hit the cap and you get auto-downgraded to a cheaper model for the rest of the week.
even if you still have "premium requests" left.
Opus models are being removed from standard Pro entirely.
github is forcing a choice: upgrade to Pro+ or radically optimize your prompts.
this is the first major platform to publicly say the economics don't work at flat rate.
i saw this coming six months ago. agentic coding isn't autocomplete. it's multi-step reasoning with self-correction and full-repo refactoring.
the token math was never going to hold.
if you're building on Copilot, you need a backup plan.
what's yours?
CEOs are facing a strategic fork in the road with ai.
cut headcount and boost the bottom line through automation.
or augment teams and grow the top line through innovation.
most are choosing door number one.
but HBR just made the case that companies picking augmentation may win long-term.
here's why:
automation is a race to the bottom. your competitor automates too. margins compress. nothing differentiates you.
augmentation compounds. ai + human judgment beats ai alone on messy, ambiguous problems.
i run a saas + agency and i've tested both.
automation saves money today.
augmentation makes money tomorrow.
the teams i'm watching scale fastest aren't replacing people.
they're giving them superpowers.
choose wisely.