Maybe we're not looking far enough up the chain. Everyone used to use forums and other websites run by other people (eg Substack) to read and discuss about things they don't understand.
Now when people have a question, they ask AI, which is wrong a lot of the time.
Visual Pinball runs on phones and tablets (and extremely well in fact) but I think the inbuilt accelerometer is not implemented. That's because during play you can't really separate movements while playing from intentional nudging very well. There are other mechanisms built into the mobile version for nudging, like tap zones.
Why didn't you time the interval between the beeps? Then when the next beep was coming up, just listen to a particular part of the room or set of boxes until you found it.
This is the problem with technology. Some people just want to use it, so it's quite hard to tell whether this is really a quality of life improvement, or if it's just exciting for some weirdo (any programmer counts as weirdo here).
I've personally had to track down a beep before and used a similar technique as you suggest: time the beeps then be in right place to perform a kind of binary search to home in on it. I found a smoke alarm in the roof space above my flat, nailed to the top of the rafters that had been there since the building was built 20 years prior. It was annoying but I'm sure it took way less time and effort than "getting AI to do it".
This is also why it's been difficult for me to decide on portfolio projects to build. I just don't have anything I really need to build. I made a compound interest calculator as my first React app, but could have made an excel sheet or python script.
I use a frustration threshold. A lot of times, I’ve got a working solution, but it’s not “right”, so when that become frustrating enough, I build something.
A few examples
dotnet does not work on OpenBSD, so to run jellyfin (supported by the infuse app on the apple tv), I have to run Linux. I wonder if I can replicate the API somehow
calibre is quite nice, but I don’t use much of its capabilities. I could try to make a simpler system for my use cases.
How much training or education is needed? I mean, frankly, a person can just go watch a 20 min YouTube video and learn all they need about how to restrict calories to lose any amount of weight they like.
>I mean, frankly, a person can just go watch a 20 min YouTube video and learn all they need about how to restrict calories to lose any amount of weight they like.
Feeling like you understand something and actually understanding it are two very different things.
Let's take something as simple as "calorie deficit."
Being at a calorie deficit itself requires knowing two things:
1. How much you burn in a day (total daily energy expenditure or TDEE)
2. How much your intake is for the day
Deficit = TDEE - Intake
You can do calorie counting for #2. But what can you do for #1?
The average guy does not know how to calculate these properly, let alone judge weight movement, which can move a lot in the opposite direction due to water weight fluctuations that can exceed 1 kg+ of movement a day.
If you calculate your maintenance calories using a TDEE calculator and get 3,000 kcal, then apply a 400 kcal deficit, your target intake becomes 2,600 kcal/day.
But what if your actual maintenance drops by 300 kcal the following week?
Your effective deficit is now just 100 kcal/day, assuming your intake stays the same.
Eventually you might find yourself barely losing weight, or not seeing any noticeable change for a while. You'll start wondering, "I'm eating in a deficit, so why am I not losing weight? Is there something genetically wrong with me?"
You might even blame your metabolism.
The problem is that your original calorie target was based on a static estimate of maintenance, not a continuously updated one. Your energy expenditure changes as your body weight, activity, and other factors change. A TDEE calculator can't account for those changes in real time.
The average guy does not understand expenditure, weight movement, the calorie-to-fat equation, water-weight fluctuations, and many other gotchas, and no, 20 minutes may not be enough to understand all this for an average Joe.
We use neural network for #1 problem, we've 17,000+ members and 2,000+ members in our telegram community group seeing great results!
if interested, you can checkout our completely free solution to this problem, https://macrocodex.app/
Guaranteed results within 2-3 weeks, no paywall, or ads.
My question is not how do you defend against it, my question is whether ablitteration makes the model completely oblivious that it shouldn't be doing 'the bad thing'. Like if you manually prefilled a normal model's output with that action, it would likely go "oh wow I can't believe I just ran that, I must warn the user of this now..."
reply