Why would this be the case when the text that produces binaries (code) is usually both more token efficient and vastly more effectively organized for modification/extension?
Unless by bits you just mean text in general, or any data since it’s all bits, in which case what you’re saying is trivially already true.
It seems like you’re saying that long term LLMs will output pure machine code as the most effective way to use them.
What about the other thing that’s the actually important part of my reply, modifiability/extensibility? You just chose the weaker of the two factors I provided. And didn’t even provide a convincing argument related to it.
You also didn’t clarify your position at all.
Explain why there is any benefit to outputting raw machine instructions compared to writing shorter, more interpretable, modifiable, extensible code and then using a compiler to turn it into machine code.
Why are agents not going to use compilers in the future?
The fastest, most surefire way to nuke your child's brain is to give them a screen.
There is certainly room for software to design a curriculum for the parent or teacher to use. We have tried (and failed) to find such a thing for our homeschool curriculum.
But giving the kid a screen? If you're even a midge honest with yourself, a screen will de-educate your child, leading to the opposite of the intended consequence.
I empathize with the screen time concerns - the reality is most kids in classrooms are already on screens, which are doing very little to move educational outcomes (from what I can tell based on the NAEP evidence).
And then you also have kids at home who are on TikTok or Instagram reels - combining screen time with easy dopamine.
Some kids using Bloomy decide to use it at home in lieu of social media time. I'd take that as a win. Cheap dopamine erodes focus and mental resilience. Bloomy focuses on ‘effortful dopamine.' We structure the experience so that students generate reward by conquering academic challenges, training the brain to find satisfaction in effort, patience, and mastery rather than passive consumption.
I'd be curious to see data that can isolate the relationship between educational outcomes and small bursts of high-quality instruction delivered via tech. That's one longer-term outcome we're trying to demonstrate. So far our efficacy results (still early pilots, but correlative) appear to be incrementally positive.
Regardless of the content on the screen, the screen time is bad for kids.
Even desiring pixel-represented "data" for this fact that is obviously observable irl is itself perhaps a result of excessive screen-time, leading to spreadsheet brain, but I can gratify you with these studies: [1], [2], [3], [4], [5]
Of course, no study will exactly disprove the claim that there is some possible content that is a net positive for kids, akin to how no study can prove that unicorns don't exist. But this is exactly the kind of brain-melt, culturally destructive kind of thinking you could expect for demanding "data" for every obvious fact.
Screen time is simply bad for kids. Don't give your kids a screen. Give them a ball, a stick, a book, etc
Just sayin