This last few days were… Different
So, for one: Last weekend I basically spent it unlocking a very-well locked laptop. More precisely, a BitLocker protected Windows would refuse to boot. The HP partition to re-install wouldn’t even see the ciphered partition. Neither would a Windows Repair or Windows Install USB.
Finally, a Live-Ubuntu pendrive gave me hopes when I saw the disk using
fdisk -lBut then, that was only the beginning. Next, I needed the BitLocker recovery password. I had to wait for a few weeks for the (anonymous, for this post) user to give it to me.
When I got it at last, I spent the weekend trying to open the ciphered partition.
Remember: Backup, Backup and… Backup
Yep. That’s like the first 3 rules of Cybersecurity.
And I should have done that in the first place: Do a “dd” of that disk the moment I saw it. But then I needed a spare 1TB disk which… I didn’t have. Fair enough: I could have bought one. I should have. I didn’t.
And so, when the time came to actually try to access the data, every single step of the way was stressful: One wrong move anywhere, I could corrupt the thing. And then forget about recovering data.
Well, that’s why I took my sweet time with it. Long hours went into this “task”. So many more that I care to admit. Oh well.
It’s done, we got the data back.
Side note
Almost all of it was already in a separate backup, so the effort was worth a measly few files… But that I didn’t know upfront… Plus, it’s been a good exercise.
I realized after the fact, Sunday afternoon, it had been about 20 years I hadn’t “dug” so deep into SysAdmin stuff.
Also: Praise to “dislocker”. Really great stuff.
And also: Ubuntu wouldn’t cut it (almost! but then no); neither did SystemRescue. Kali was the distro that would have the up-to-date-enough version of dislocker in the end. Without which, there was no opening the f$*#” disk… (A few hours lost right there…)
So what’s next?
Well, a few other things happened.
I have a presentation ready… To do a 1h+ stint. It turns out, I will only get 15’. I should have asked earlier, this one is on me.
So first, I need to make my RLCS presentation comprehensible and useful for a 15’ (tops!) session. Which, for who knows me, is difficult. I have a much easier time talking more, than I have talking less!
“It’s long because I didn’t have the time to make it shorter” is a saying (no idea of the source). It applies, here. I have a month ahead of me to squeeze the valuable stuff of the RLCS package into a 10 to 15’ presentation. Alright, it’s not that bad, feasible.
And then, for no reason whatsoever, I got reminded (thanks random LinkedIn posts) about the ARC-AGI1 challenge. And it got my head spinning.
ARC-AGI as a new Blog topic
Well, it’s ambitious indeed. But heck, why not? So it’s official, I’ll look into it. Even if I finally decide I can’t barely start tackling the challenge, I’m somehow certain I will learn stuff along the way.
Heck, a few initial outcomes of simply thinking about it:
After a few hours giving it some thought, it’s clear almost none of what I have done thus far will help me with this challenge. (Link in resources, if you don’t know it).
I think one key aspect of this “few shots learning” approach, after trying some examples myself, is about finding what changes. So I’m thinking XOR right there. But XOR what?
I think I need to somehow… Represent knowledge? IDK! Oh, and this reminisces of course some “Knowledge Graphs” concepts, which I was not a big fan of, but maybe now I’ll give them a thought.
Talking about graphs!
One aspect of the ARC AGI thing that isn’t helpful is… It’s a rather “visual” challenge, and I’m more of a tabular-data or, sometimes, text-only-data person. So this is a challenge a bit. But I already learnt I can do “kernels” and understand them a bit…
Well, I found something else that is interesting for this challenge, and uses graphs! It’s an algorithm called “Connected Components Labelling” and I found it marvelous, well because it uses graphs :)
I won’t go into details, but it turns out it’s just so very similar in my mind to the Bongard problems (which I discovered in the “bible” GEB…)
Various notes
I have yet to finish Hofstadter’s “Gödel, Escher, Bach”. I just don’t fathom reading it with anything less than 100% of attention, and no more than one chapter at a time. And I read quite a few books at the same time, and so this one is taking me forever! But I definitely intend to finish it in due time.
Oh, and yes, there is an ARC AGI2. But let’s not even get there yet. It’s not the goal.
Conclusion
For my anonymous person, good results: We got her data back.
For me? Not so tangibly-productive a week, I guess, but I have had time to think about innovation (oh, that’s right: I’m also reading “The Idea Factory”, about Bell Labs, by J. Gertner, and enjoying it inmensely, although here again, I’m taking my sweet time, maybe getting a couple of chapters out of the way each week… Still!). Right, so, thinking about ARC AGI, innovation, but also about Cybersecurity, as I have drifted a bit away from my “core topic” by following my various curiosities…
Sometimes, not doing, thinking, is good too.
Resources
http://github.com/Aorimn/dislocker -> VERY USEFUL.
https://arcprize.org/arc-agi/1/ -> VERY INTERESTING.
https://en.wikipedia.org/wiki/Bongard_problem -> Maybe inspiration for Chollet’s ARC AGI?
https://en.wikipedia.org/wiki/G%C3%B6del%2C_Escher%2C_Bach -> THE “GEB”, no two books like this one.
https://en.wikipedia.org/wiki/Connected-component_labeling -> I definitely will take this as part of my work on the ARC AGI challenge!