Internet is becoming more and more polluted with junk-mail, people selling crap, and businesses which don't know their place on the net. They're all trying to make this wonderful place (i.e.: the net) in to hell (i.e.: real world). Internet should be viewed as a place of imagination, creativity, and most of all: fun. Internet is not some really advanced tool for searching for people to rip-off. It's about searching, and finding, things which are useful, helpful, and promote the sharing of ideas. This is what this site is striving to become.

Really good paper: Universal and Transferable Adversarial Attacks on Aligned Language Models

- Alex; 20230815
August 15th at wikipedia...

Finally got the bigger (llama-2-70b-chat) Meta's Llama 2 model running locally on my desktop via llama.cpp. It's quite slow, but it's quite amazing to have something similar to chatgpt running locally.

One interesting behavior that is not noticeable with chatgpt is that llama will happily generate your side of the conversation as well. You ask it a question, and it responds. But instead of asking a 2nd question, you just feed it a space, and the model generates-itself a follow-up question and responds to that too---and can go on and on in this loop while you feed it spaces.

- Alex; 20230731

It seems Elon Musk has formed xAI, with the goal of ``to understand the true nature of the universe''.

Why do folks think that AI will be any good at this? For one, the true nature of the universe may actually be uknowable (our knowledge is limited to observations we make---we can never see the true universe---just the photons that we record with our instruments).

Similarly, if Bible was part of the training data, then the true-nature-of-the-universe may have an easy and simple explanation :-)

- Alex; 20230712

There is a concept of a force multiplier, which refers to something that gives us the ability to accomplish greater things than without it. Mechanical tools, such as hammers, let us multiply the effect of physical force.

Sometime in 1980s Steve Jobs likened the computer to a ``bicycle for the mind.''

Computers are tools that amplify effort in tasks we generally associate with mental activity, such as memorization, recall, and calculation.

There is a certain limit in that analogy: the mind is still the mover that makes the machine go---bicycles make legs more efficient, but their output is limited by the output of legs.

The recent intelligence engines are on a whole new level: they are like motorcycles for the mind---the mind does not have to make them go---and their output is not limited by the mind.

In fact, because these engines can consume more data than any individual human, they are already a lot ``smarter'' than any individual.

Unlike humans, these AIs can share what they learn: a thousand instances can learn things in parallel, and aggregate all that knowledge within each engine. They can also retain knowledge forever: they don't die and take their experiences with them. Imagine a doctor who has observed billions of patients over many generations---they are bound to spot patterns that human doctors would miss.

It seems intelligence is not special. Evolutionary speaking, our legs are not the fastest way to get around the planet [motorcycles, cars, trains, etc.] A bird's wings are not the fastest way of flying [jet engines, rockets, etc.] Our eyes are not the best vision instruments [telescopes, microscopes, etc.]. ...and our human intelligence is not the best intelligence... we are not the center of the universe...

The digital brains appear to pack information better than biological brains---and according to Geoffrey Hinton, the backpropagation learning algorithm may even be optimal [in some sense]---so digital brains are better at learning.

- Alex; 20230614

Re-reading The Grand Design by by Stephen Hawking. Last time (October 2010) I didn't think much of it. E.g. M-theory and multiple-descriptions of everything coupled with anthropic principle. More recently, found a few articles regarding Model-dependent realism and Stephen Hawking's Philosophy: Model-Dependent Realism that resonate with many ideas I've been having.

Last summer, I was re-re-re-reading Quantum Mechanics (The Theoretical Minimum), and during one of my hiking trips, something obvious occured to me: the rest of sciences are just like quantum mechanics, we're just not accustomed to viewing them that way---that is not to say quantum mechanics rules over all other sciences---but the way quantum mechanics views reality should be applied everywhere else.

For example, what we think of as equations are actually relationships (that's the obvious part): for example, F=ma (and other formulas) can be rewritten as (F,m,a) tuple (relationship) such that |ma - F| is 0, but in any actual experiment we can never observe that exact 0... so all (F,m,a) tuples that get close to zero are "possible" observations from experiments.

Since this eliminates ``functions'' [pure definition of a function: y=f(x), cannot have more than one y for a given x], it also opens up the door for multiple solutions to things... (e.g. quantum mechanics): suppose our measurement tuple is (a,b,c,d), and we observe values of a,b there's a distribution of values for c,d that would be possible from other observations (e.g. each of a,b,c,d are "observations"---there are no independent or dependent variables... they're all "observations" that we are making; we could just as easily observe b,c,d, and try to figure out the distribution of possible values for a.) Kinda doing marginalization. This implies that distribution for variable a may be multimodal given some observations, and unimodal given other observations [and making an observation of another variable "collapses" our knowledge, and there's no need for faster-than-light travel or spooky-action-at-a-distance since... well... the only thing that collapses is our knowledge, etc.].

The other obvious thing that occured to me is that these are just structures to model what we observe, not how things are. e.g. Newton's laws worked just fine to model what we observe---but lots of folks took it to mean that planets move due to the force of gravity, etc. But later (relativity) it turned out that they move due to curved space-time (not "force" of gravity)... but that's just another construct just like newton's laws. e.g. Newton's laws had some observation tuples e.g. (a,b,c,d,e), etc., and worked fine within that concept-space (given observations of a,b,c,d, we could marginalize out and guess at values of e). Relativity does the same thing, except the tuples are different... we know that given observations we can guess at values of other observations... but relativity itself doesn't actually tell us the actual mechanism (it only tells us what we are likely to observe). Same for quantum mechanics.

So that means all sciences are just surfaces/curves in n-dimensional space where some quantity is minimized. e.g. newton laws are essentially a surface in 3D where (F, m, a) where |F - ma| is close to zero. Every point in that space will have a probability of being observed... with points close to the surface having much higher probability. My guess that's mostly what the principle of least-action is... but I never seen it stated that broadly. It is important to note that some of these elements of that tuple may be measured/observed, while others are our own inventions... for example in (F,m,a), we may measure the mass, but we have to estimate acceleration, and F is a completely imaginary entity... but it's part of Newton's model.

In other words, some things in the model may be observables, others are things we can caclulate from observables, or use to build a distribution of observables before we make an observation.

Now for the reason to re-read The Grand Design: There's an implication that we can have multiple systems that are not compatible with each other, yet both work to predict observations. e.g. relativity vs quantum mechanics. Within the realm of relativity certain observation tuples will have higher probability of actually being observed, and that's fine. Within the realm of quantum mechanics certain observation tuples will have a higher probability of actually being observed, and that's fine too. They can coexist---since both of them are based on what-may-be-observed, there's no inconsistency. (inconsistency arises when we start to assume that the theory implies how things really are, not just what can be observed).

In other words, imagine general relativity model as tuple (a,b,c,d,e,f,g), while quantum mechanics as being (e,f,g,h,i,j,k,l,m). There is some overlap---the obesrvables/elements (e,f,g) may be applicable to both theories---and that's how we can tie them together. BUT, going from (a,b) to (k,l,m) via (e,f,g) may not work quite right. We may have a relationship, but the distribution of possible observations may make the result meaningless. (e.g. there is noise in each measurement/observation---so very precise measurements in quantum mechanical realm may not give any meaningful predictions in the relativity realm). We may have two ``contraditory'' theories that both work. Model-Dependent Realism indeed!

Imagine we have a nail half stuck in some surface. We shine a light on it. It casts a shadow. We have (shadow, light, nail) as our model [shadow=shadow length, light=height of light source, nail=height of nail]. None of these cause the other. Before measuring, there's a distribution P(shadow, light, nail) of all possible values... suppose we measure the shadow (using a ruler, so some measurement fuzziness is present in the measurement), then we have P(light, nail | shadow=X ) where X is our measurement... so now we have a distribution but with less uncertainty... then suppose we observe nail height (again, using a ruler), and we get P(light | shadow=X, nail=Y)... and again, we have a distribution, but with a lot less uncertainty. There's no "formula" to compute height of the light, the prediction from our observations is just a distribution of how likely we are to observe a particular value of light.

Now, we could "model" that nail, shadow, light relationship with an equation, that's like figuring out the parameterization for a distribution... OR... we could model that relationship [manifold in 3-dimensional light,nail,shadow space] using kNN or something (e.g. make lots and lots of observations of triplets (nail, shadow, light), and for any future observation we could just pick the k-nearest-neighbours.... either way we get a set of possible values for what we may observe next. So the distribution of observation values is primary (that's what we observe), and our laws are like regression of those observations into equation form---and there may be different ways of "regressing" into equation form (e.g. same observations lead to Newton laws, and same observations can lead to Relativity).

There is also an interesting property that we can project that ``measurement tuple'' into smaller dimensions. For example, if our ``theory of everything'' has (a,b,c,d,e,f,g,h,i,j,k) variables, we can project that onto say 3D as (b,f,k) and only work with those variables, see what theories we can build on top of those. For example, relativity is one such projection, quantum mechanics is another such projection, etc.

Similarily, quantum uncertainty pops out by itself: The tuples would have related variables. Suppose we cannot measure two attributes at the same instant, nessasitating two measurements, so we would have (time1, attribute1) and (time2, attribute2), now we may construct (time1, attribute1, time2, attribute2), and filter based on proximity of time1 and time2, and project onto (attribute1, attribute2)... but that's our derivation---we are assuming that these observations were made ``at the same time''---while they may not have been (or may not even be possible to make). So everyday human-sized objects, we'd use (attribute1, attribute2) without issues, e.g. F=ma style physics, while in more precise terms we may use (time1, attribute1, time2, attribute2) style physics, which will have a distribution for every attribute.

UPDATE: Regarding sources of measurement error/noise: inverse square law implies less and less information reaching the destination. With noise present, we can calculate how many bits of information will reach---we don't know the noise level though, but, we do know that signals are quantized (particle detected vs not), at some point the signal will completely disappear and all we would be detecting is noise. What this means is that due to noise, two objects in the universe can be completely cut off from each other---this could be the reason why the universe is expanding---since more and more of the universe becomes gravitationally detached from each other.

- Alex; 20230511

Geoffrey Hinton quits google, so he could speak freely about the dangers of his life's work. From the article: ``Hinton was happy with Google's stewardship of the technology until Microsoft launched the new OpenAI-infused Bing, challenging Google's core business and sparking a "code red" response inside the search giant. Such fierce competition might be impossible to stop, Hinton says, resulting in a world with so much fake imagery and text that nobody will be able to tell "what is true anymore."''

It echoes my blog post from March 24 :-D

- Alex; 20230430

I have this weird feeling that the internet is about to be overrun with bot-generated content. Kind of like we can pollute the environment, I suspect the internet, far from being perfectly "clean" as it is, is about to be drenched in crap---so much so that generated crap will likely outnumber anything non-crap by a huge fraction. Perhaps a million to one. E.g. try finding something useful, when content matching your search criteria is overwhelmingly dreamed-up/hallucinated by chatgpt or similar.

The Kindle Books spam is just the start... wait till the entire planet realizes you can build anything... websites, games, etc., just by asking a bot to do it for you. They all will, and it will all find a place on the internet. The internet's signal-to-noise ratio will drop---as low as it is now, imagine it being almost zero.

...and then Phase 2: bots/chatgpt will consume this toxic sludge as input, and inevitably produce more crap---crap feeding on crap. Perhaps the only way to find any useful/factual info in the future will be to use archived internet from 2022, before the bots started passing it through their digestive track over and over and over and over again... kinda like picking up a Kindle-Direct-Published book after 2022 is kinda pointless.

- Alex; 20230324

Our Hoverboard just died. Hover-1 Rogue... it just completely died and refuses to charge :-/

- Alex; 20230323

Skidded off into mud on the scooter. More hurts all over :-/

- Alex; 20230313

It seems more things got hurt. Now elboes and shoulders are hurting too :-/

- Alex; 20230221

re-Learning how to ride a skateboard. I had one when I was a kid, but haven't touched it in many years. Ended up planting my face into the pavement. Bloodied chin, knee, and arm :-/

- Alex; 20230220

Visiting The Steven F. Udvar-Hazy Center. The SR-71 is one amazing airplane.

- Alex; 20230219

Doing a road-trip to Virginia Beach. Plan for the long weekend: learn to ride a skateboard.

- Alex; 20230217

Interesting: Nvidia boss Jensen Huang calls ChatGPT the greatest thing to ever happen to the computing industry. ``an "iPhone moment" for artificial intelligence''.

Is it on the same scale as a steam engine or invention of Excel? It will certainly change some industries.

- Alex; 20230214

Interesting stuff: Black Hole Star - The Star That Shouldn't Exist. Would certainly explain the origin of super-massive black holes at centers of galaxies.

- Alex; 20230213

Uh, oh! Alphabet Stock Loses $100 Billion After New AI Chatbot Gives Wrong Answer In Ad.

- Alex; 20230208

Uh, oh! Microsoft announces new Bing and Edge browser powered by upgraded ChatGPT AI.

- Alex; 20230207

Happy New Year!

- Alex; 20230101

Have been playing with ChatGPT for a bit, and so far, it's very impressive. Besides using it to write letters of recommendation, compose questions for an exam, and, probably answer exam questions, it can also create new stories on the spot!

For example: "write a story similar to little red riding hood, but in a way that makes me want to drink a diet coke."

Once upon a time, in a bustling city, there lived a young woman named Cherry. She was known for her bright red hoodie and her love for diet coke. One day, she decided to visit her grandma who lived on the outskirts of the city.

As she walked down the busy street, sipping her diet coke, she felt a sudden thirst for another one. She stopped at a vending machine and got herself another can of ice-cold diet coke. As she continued her journey, she met a sly wolf who tried to trick her into telling him where her grandma lived.

But Cherry was smart and didn't fall for his tricks. Instead, she offered him a sip of her diet coke and told him that she was on her way to see her grandma. The wolf was taken aback by her kindness and decided to join her on her journey.

Together, they walked through the lush forest, chatting and sipping on their diet cokes. As they reached the cottage, the wolf got the shock of his life when he saw that Cherry's grandma was not the frail old lady that he had expected, but a fierce and independent woman who was a huge fan of diet coke too.

The three of them spent the day chatting and drinking diet coke, and the wolf even helped them with the chores. As the sun began to set, Cherry and her grandma offered the wolf a bed for the night and even gave him a can of diet coke before he drifted off to sleep.

The next morning, the wolf left, a changed man, with a newfound love for diet coke and the kindness of people like Cherry and her grandma. And Cherry and her grandma lived happily ever after, always keeping a steady supply of diet coke in their fridge.

The end.

Feeling thirsty yet? Grab yourself a diet coke!

- Alex; 20221215

Uh, oh! OpenAI's new chatbot can explain code and write sitcom scripts but is still easily tricked.

- Alex; 20221201

Been thinking about entropy vs time: entropy always increases with time. Second law of thermodynamics.

Suppose all we have is an analog watch---the circular dial 12-hour kind. Nothing else except this. Can we measure time intervals longer than 12 hours? Technically, we can, by remembering how many times the watch returns back to the same spot---but what if all we have is that analog watch---then every 12 hours, we are back to where we started. Suppose we measure entropy using such a clock---the amount of disorder would tend to increase for 12 hours, but then reset (as time resets---there is no memory of disorder prior to the start of our measuring).

But yeah, we could remember how many times the hour hand goes around the circle, but that requires another counter somewhere (outside the watch). Suppose we build such a counter that can count to infinity---starting at some arbitrary moment which we will call time zero. The longer times goes on, the more memory this counter will consume (it will eat up logarithmic number of bits---slow, but still consuming bits of information). The act of measuring time seems to itself increase entropy (those bits used to run the counter).

What if this is what entropy is? E.g. instead of us measuring the increase in entropy, our measuring is what causes us to perceive the increase in entropy? For example, if we lived in a universe where the most precise clock we could have was a 12-hour analog watch, then entropy would not always-increase---but because we have a device with a bigger counter (e.g. the universe may be such a device), we observe it counting away, and the counter is always incremented---just like entropy. So entropy always-increasing could be an artifact of the way we make observations and measure time.

- Alex; 20221121

...and back in NYC :-)

- Alex; 20221024

Arrived in Phoenix, AZ. Got supplies, and heading to Grand Canyon.

At Grand Canyon (south rim), clicked some pictures of the milky way. It is a moonless night, and stars are amazing.

Around 8am, headed down via South Kaibab trail. Weather is amazing. Sunny, but not too hot.

Around afternoon time, weather started to get cloudy. Reached the grand canyon north rim a bit before sunset, and just barely made it to the Bright Angel point before actual sunset---it's about two mile walk.

Weather got very windy (gusts like seem to almost knock ya off yer feet), very cloudy, with drizzling rain. Couldn't chill out at the visitor center due to wind. So headed down the north kaibab trail.

Lots of hikers reaching late.

Couldn't chill out at ranger station, nor Cottonwood, nor Phantom ranch---the wind was too weird, it seemed to be ripping trees right out of the ground.

Headed up the south Kaibab trail around 2am. At Tip-Off point got pinned down by wind and hail . It's a roof, but no walls. In summer it's a perfect place to nap for a few hours before the long hike---but ended up shivering and wet instead. After about two hours of just sitting and shivering, the wind quieted down. So headed up the trail.

It got slightly brigher. There's ``snow'' (hail) on the trail. Got a bit too exhausted this time around due to lack of any place to rest---barely walked up the south Kaibab trail. Caught a shuttle bus at the top.

McDonalds for breakfast, then to Walnut Canyon, then to Montezuma's Castle (which turned out to be a hole-in-a-wall-house).

Then due to traffic, was late to the airport, and myraculously made it just barely on time. Had a few minutes before boarding, so set alarm to wake me up in 20 minutes... and next thing I know they're doing last boarding call for me... so woke up and ran into the airplane.

So yey, fun weekend!

- Alex; 20221022

In ther news... try: Rewind-Replay.

- Alex; 20220928

Kiddo's class issue resolved :-)

- Alex; 20220909

First day of kiddo's school. Something had to happen on day-1... and it seems kiddo ended up in the wrong class.

- Alex; 20220908

Got around to buying DOOM Classic Bundle on Steam (it's around $5 at the moment). It includes DOOM 1,2,3. Lots of love for that game, as it inspired me to get into programming. I held the floppies in my hand, and pictured there's a whole virtual world in-there on those floppies---and I just gotta know how to create these things.

- Alex; 20220823

...and back in NY.

- Alex; 20220822

Visiting Shimla, and Jakhu Temple. Kids loved walking up the steep road/stairs.

Shimla has strange architecture: The hotel lobby was on the 7th floor, while we stayed on 2nd floor, and car dropped us off below 1st floor (kids ran up the stairs all the way up to the lobby).

- Alex; 20220816

Flying out to India.

- Alex; 20220812

5-hours of John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets :-)

- Alex; 20220810

There's no way the boots will dry by morning: what hike can I do in wet boots? ...and then it occurred to me... I could go to Zion and do Narrows! So 8 hours of driving later, I entered Zion and was on one of the first buses out.

After being chilled the previous day, decided to do a warm up hike (before, submerging myself in cold river water). Wanted to do Angels Landing, but it seems you need a permit these days: which is easy to get, you just can't get it on the same day... wish I thought of this earlier.

Anyways, near Angel's Landing there's a ``west rim trail'' that goes up up up and along the west rim of the canyon. So ran up that one for about an hour in and an hour out. It was amazingly nice. I never done it before, and will not skip it from now on... Angel's Landing is nice, but west-rim-trail is also amazingly good.

After that warm up, with the sun mostly out and warm, headed for The Narrows. Often I ended up in Zion during odd-times of the year, so Narrows would've been too cold, and I didn't realize that you didn't need any equipment to actually do it until recently. Anyways, the narrows river is mostly shallow... maybe a foot deep in most places. There are deeper spots that go upto the waist, and some even deeper spots (that you avoid by checking depth with a stick).

About a mile and half in, there was a rock right in the middle, with very deep water on both ends. The few folks I saw who tried to cross it went as far as their necks. That's a bit too deep for my first attempt, so that's the place I turned around.

The whole hike was amazing. It was very fun. It's hard to explain how awesome it was. Most fun hike ever. Will definitely come back to this again (and again, and again, etc.). The weather was just right for it... the water was cold, but not too cold... the sun made everything warm, but the water kept everything cool. It was just perfect. One thing for next time: get proper sturdy walking sticks. I had to improvise with drift-wood, and those sticks break when you put weight on them (which happens often due to water current and shifty rocks under the water).

On the way out of Zion, did the Observation Point hike that I've never done before. Then headed to Salt Lake City airport.

- Alex; 20220807

After Yellowstone, headed towards Grand Teton National Park. Last time I tried to do Static Peak was in June 2016, and my words from then: ``First snow started in patches, then more patches, then sections that had to be traversed, then it was just endless snow. So about a mile from the summit, it was just ridiculously suicidal to continue... the slope was just too much, and sliding down would mean a pretty long fall off the mountain... so decided to turn back without reaching Static Peak''

This time I'm back to finish that hike! So got to trailhead, and headed up the Death Canyon. There didn't seem to be any snow, but it was foggy. Then it started raining. Then it started raining heavily. A few hours into the hike, I was totally soaked. It was hot and humid, so rain was pleasant.

Then altitude started to make the wind colder. Rain got much colder. My wet-boots/shirt/pants/underwear started to get chilly: it is all synthetic, so less danger of hypothermia, but still gotta generate enough heat to keep warm. Put on jacket, and continued walking up.

Got to the point where I turned around in 2016. It was a bit less than a mile from the ``summit''. Now for the confusion: there's no summit. The trail is to ``Static Peak Divide'' (not to Static Peak summit). So you get to the static peak divide, which is just a sign saying you're there. I thought that the summit is just a bit farther up, so headed farther up. The trail does indeed go a bit up, and then it goes down (the other side of the mountain, without traversing the summit). Anyways, I kept on walking until I reached a sign that says you're at the boundary of the national park. So that's where I turned back: didn't wanna continue down the mountain into a national forest, which leads who knows where. So yeah, there's a ``Static Peak'' mountain, but the trail doesn't take you to the summit [and it doesn't look like a summit you could just walk up... very pointy and steep up there]. On the way there were a few patches of snow, and I slipped on some muddy hill :-/

By the time I turned around, I was actively shivering from the cold, and lost sensations in my hands and toes. Feet were wet for at least 5 hours by then, and the cold rain didn't help things. Tried running to warm up---didn't really help. The problem is that any heat you generate is almost immediately lost to freezing rain. Anyways, at a steady pace, got out of the wind, then rain got a bit less, and slowly but surely, by the time I was nearly by the parking lot, the rain has stopped and was replaced by humid sunshine.

This was the most miserable 10-hour hike I've ever done. Trail was foggy most of the day, it was very cold, windy and wet, and I didn't get to the summit that I've been waiting to get to since 2016! (ok, that last part was due to poor research/planning).

...and my hiking boots got wet :-/

- Alex; 20220806

Arrived in Salt Lake City, rented a Jeep, and headed north. Found a quiet rest area and spent most of the day working. Towards the end of the day headed for Yellowstone National Park.

Wanted to explore a ``small'' hike near Old Faithful: Observation Point trail, and that trail that loops around Old Faithful. It starts/ends at old faithful---just go right around old-faithful and follow it along boardwalks---it's a few miles long, and goes through a lot of active geysers.

It seems it didn't get fully dark yet, so headed for Prismatic geyser, which was packed with visitors (had to park outside the main lot, and walk down).

Then headed for Artist Point---so managed to hit the 3-primary attractions in Yellowstone in just one evening... and then it got dark. Since campsites at Yellowstone fill up quick, improvised and "car" camped by some construction equipment on the side of the road (e.g. nobody notices some cars parked by a bunch of excavators). Good thing I brought a sleeping bag, as it got pretty cold there at night.

- Alex; 20220805

...and back in NY, just in time for a dentist appointment :-/

- Alex; 20220801

Towards morning, the Homestake Dry Camp got a bit chilly, and I had to hop inside the truck to warm up. Woke up around 6am, and decided not to waste day-hours and get going early.

Went for a walk on the Racetrack Playa to find sailing stones. They are all the way in the back (by the mountains) of the south-end of the playa. Found a few---but not as many as before. It seems the best time to see sailing stones was a decade ago :-/

Then just walked around the playa. It is an amazing place to just walk and walk and walk and walk.

It started drizzling, and temperature was not terrible. So decided to visit Badwater again.

At Badwater, the sky was overcast, and it seemed like it just rained. Walked around a bit, then headed back to Vegas.

It started drizzling a bit. Not really raining, but a bit of rain drops here and there. Somehow that was enough to cause a flash flood. Actually, there were several flash floods. With phone beeping that national-emergency message to not travel. The truck I was in drove through most puddles without issues.

But then there was a buddle with a pickup truck, and an amazon delivery van stuck in the middle of a road-that-turned-into-a-river. There was a huge queue of cars waiting for something. I drove onto the side of the road to get a better view of the puddle.

A tesla attempted to cross the puddle, and made it! Then another tesla crossed just fine. The water went almost to the roof! These things really can drive underwater!

A few minutes later, a pickup truck (similar to the one that was stuck) decided to cross, and it made it through! My guess when the flood initially happened the water was higher.

So I decided (in a Toyota Tacoma with all-terrain-tires in 4wd mode) to drive through the puddle as well: The water went over the hood in the Tacoma, but it still kept on going. I know that air-intake is a bit over-the-right-front-wheel, but somehow that did no get clogged with water, and the Tacoma kept on going. I made it through. There were a few other sedans stuck on the other end of the puddle and a similarly long queue of cars waiting for something.

In retrospect, it was stupid to try to cross that. It was nice that it worked out though, but it could have ended much worse. Stupidly (or smartly?) I kept windows open (just in case I got flooded and needed to get out quick), and the muddy water made it into the car. Stuff inside car got a bit dirty, and I had to clean it afterwards: good thing I had 4 galons of water left after the trip.

In the next town after the puddle, saw a firetruck heading down to that flooded road. Anyways, made it to Vegas. Parked in Vancian (free parking), and walked around the strip. Walked all the way to Luxor and back, and then it was time to return the car and head to the airport.

...and then the flight was delayed by a few hours :-/

- Alex; 20220731

At Mahogany Flat Campground the temperature was 60-ish. It's at 8k feet above sea level---so much cooler than the last campground (which was around 100-ish degrees, even at night). Anyways, the sky was clear-er, so did some stargazing in the back of the truck. The milky way is very clearly visible, saw 8 shooting stars, a satellite, etc. It was amazing. This is exactly what I came to Death Valley for (and exactly the reason I got a pickup truck).

An hour later woke up because sleeping on metal-in-the-back-of-the-truck at 8k feet is cold. So got into the truck and slept on the rear seats.

Woke up around 6am, and headed up Telescope Peak trail. It's a 7-mile hike, that alternates between very-steep and very-flat. It's a very nice walk. About 3-hours later got to summit: and there's an actual reflector telescope at the summit (someone must've left it there). There's also a pair of binoculars. On the trail there were a bunch of trees that were hit by lightning---you could tell they were very weirdly split/burned.

On the way down the mountain, visited Wildrose Charcoal Kilns. These seem to have been built from local rock, and if you walk around them, you can get an idea of how they were built---there's a way to climb them, etc.

Also on the way down, spotted a hundred or so ``wild'' (?) donkeys (they didn't have any markings on them to indicate they weren't wild; they probably weren't wild, but they were just out and about with nobody around). Also passed more ghost towns/settlements, and got out of car to explore a bunch. Found an old car that must've been flipped a hundred years ago.

Went to explore the Mosaic Canyon. By then it was probably 115-degrees or so, and it was a terrible idea... but I still went for a walk. About 20-minutes later, I barely got back to the car; with sweat literally blocking my eyes.

After this, the sanest thing to do was to go to Badwater (even hotter place), and walk around there... it was 118f at visitor center, at badwater it was probably around 120f. I decided to walk/run out for 10-minutes, and walk back 10-minutes. Total trip probably took 30 minutes due to walking-back-time. But yeah, walking in such heat is not easy. I did find salt-flats, but phone got the "phone overheating" warning and stopped working: so didn't click any pictures.

Decided to cool off at Dante's Peak. This is around 10k feet elevation right above Badwater. It's a long drive, but it's also much cooler due to elevation. Chilled out there until almost sunset. Then headed to Homestake Dry Camp; that's another free campground right next to the Racetrack Playa.

The Racetrack road got improved a lot. They flattened it out, and removed all the big rocks. Now it feels like a well maintained dirt road--similar to Titus canyon or Mosaic Canyon road. I'm sure a corolla would make it down that road (not that I suggest trying that, just saying there's nothing technical about that road anymore).

Got to the campsite when it was dark. Sky again overcast, and temperature around 80-ish degrees. Got in the back of the pickup truck and slept looking at the sky. About an hour later, the sky cleared up and I could see the milky way. It was awesome.

- Alex; 20220730

Got to Vegas, and started drive to Death Valley. Rented a stock Toyota Tacoma. On the way got supplies for the weekend (8 galons of water, two pizzas, potato chips, cookies, two coffees, etc.).

Right out of Vegas there was a sand-storm. The sky got dark, and just like a thunderstorm (there was lightning), there was a sand-storm---visibility was very low, and wind very high... but no rain at all. Never seen that before. Somewhere else (not far away) it appears to have been raining, with flashfloods in Vegas, etc.,

Driving into Death Valley via Titus Canyon road. Last time we drove it, it had some `interesting' bits, this time they must've improved it. It was fairly safe-seeming drive. Much of it is still not wide enough for two cars, but for a one-way single-car-lane road it didn't seem dangerous at all. Stopped over at Leadfield (a ghost town that's right on the Titus canyon road); walked around the ruins: very creepy when you're the only one there.

After Titus canyon went to Death Valley Furnace Creek visitor center. Got map, figured out free camping sites. Headed to Badwater, thinking that it might be a good place to hang out and eat pizza in the evening. It was dark. Like very dark. The sky was overcast, no stars, no moon, etc., once eyes adjusted to the dark, even then couldn't really see anything. But eh, sat in the back of the Tacoma and ate pizza. Then drove to Emigrant Campground (free).

There was one other car at Emigrant Campground. It was 100-something degrees, so I just lied down in the back of the Tacoma looking at stars. I should've brought a sleeping pad or something soft---the metal wasn't very comfortable, but it was very warm. The bugs weren't a problem once I turned off the light (I thought they'd drive me crazy, but once the light was off, there were no bugs at all). Anyways, there were no stars (still overcast), so just closed eyes and passed out.

Few hours later woke up due to: either being too hot, or being too uncomfortable sleeping on metal pickup-truck bed. The sleeping in the back of a pickup truck is awesome, except need a sleeping pad and a pillow (and maybe a blanket). Anyways, decided to drive a bit closer to the Telescope Peak trailhead, and a bit higher up: Mahogany Flat Campground at 8k feet elevation (also free campground).

- Alex; 20220729

Google Engineer Fired After Claiming AI Chatbot Had Become Sentient. Well, I doubt he was fired for claiming anything---he was probably fired for going public with non-public info: it's not his job to determine what gets released to the news.

That said, that conversation from the article does look interesting. Certainly more advanced than Prof.Phreak. Granted core of profphreak was written 26 years ago (time flies, eh?).

- Alex; 20220725

Hiking Mnt.Marcy. Doing a loop from Adirondack Loj to Avalanche lake, then to Colden Dam, then to Marcy summit, and back to the Loj. This loop took almost 10 hours.

- Alex; 20220723

Resubscribed to World of Warcraft. Might as well get to the max level before the next expansion pack. Also pre-ordered next expansion pack. Yeah, this game is old, addictive, ``boring'', etc., but it's also ridiculously relaxing.

- Alex; 20220721

...and back in NY.

- Alex; 20220720

Flying home to NYC.

- Alex; 20220719

Decided to come back home early. Adjusted car rental. Changed airline tickets. Driving back to SFO.

- Alex; 20220718

Got to North Cascades National Park. Drove to Hidden Lake Trailhead, which turned out to be a very narrow twisty forrest dirt-road. The huge truck barely fit in some places: good thing only had to pass someone once: most of that road was way too narrow for cars to pass each other.

The hike was amazing! It starts out in the forest, with a typical forest trail, but then opens up into a green valley full of vegetation (reminded me of Kalalau hike in hawaii: very steep, walking on side of cliff, with all sorts of dense vegetation all over). Then the hike goes up the side of a mountain that's still covered in snow. About half the hike is walking on a side of a snowy mountain. (wish I had microspikes). The ``summit'' (not really the summit, but an observation point) is all under snow. Got to see the Hidden Lake, but it was all hidden under ice---with thick fog for good measure.

There's a cabin there as well, but it's impossible to see tails under snow, and after looking for it (in fog) for about half an hour, decided to give up on the cabin and turn back. Don't wanna get lost on a strange snowy mountain, in fog, at night.

On way down, it's much easier to ignore the trail (as many folks have done), and just slide on shoes or other parts of body down the snow... I tried to slide on my hat, but that wasn't slippery enough. Very fun descent though :-)

All in all, this hike is awesome. Definitely best hike in Washington state. Mnt.Rainier hike is also pretty good, but this one was way better IMHO. Need to figure out how to come back and do it on a sunny day.

- Alex; 20220717

Rented a tickup truck (dodge ram 2500 heavy duty)... and it's huge. With california gas prices, I expect gas will be the biggest expense on this road trip.

Bought a cooler, and stashed it in the back of the pickup truck. Then after a bit of driving, noticed it fly out of the back of the truck! Stuff doesn't just stay in the back... Ended up buying anoter cooler :-/

Anyways, reached Mnt.Shasta. Plan was to hike to the summit, but the park ranger said that going by Avalance Gulch trail this time of year is a terrible idea, so shelved that plan---hiked to the Helen Lake (highest ``safe'' point via Avalance Gulch trail). The `lake' is a frozen mountain of snow, so got to make snow angels in July :-) Definitely want to come back to summit this mountain: it's one amazing hike.

After that hike, drove to Crater Lake National Park, and did some sight seeing (the watchman trail was still closed due to snow).

- Alex; 20220716

Flying to SFO, will drive to North Cascades over the next few days, and hike the Hidden Lake trail on Sunday. (at least that's the plan).

- Alex; 20220715

Scheduling a trip to North Cascades for the weeend.

- Alex; 20220712

...and back in NYC.

- Alex; 20220711

Happy b-day!

In ther news... around west Illinois, was pulled over for ``speeding''. Eh?

I'm the most boring driver in the universe! Literally! I was once pulled over in LA for driving too slow (going 50 in 65 zone). Here's my driving style: I set cruise-control to 5-miles (and often 10) below the speed limit. Eg. if limit is 65, I generally drive 55, if limit is 75, I drive 65, etc.

The limit was 70, and I was steadily going at 65 in the slow lane---not passing anyone, etc., this is way slower than the rest of the traffic, which was generally moving 75-80ish.

So then I see this cop car behind me, going at the same pace as me. A few seconds later, they turn on lights and pull me over. ``Do you know why I pulled you over?'' ``No''. ``You were going 74 in a 70-mile zone, while overtaking someone.'' ``You gotta be kidding me, I got my cruise control set to 65 for the last hour, and I can't remember last time I passed anyone.'' ``The radar says 74.'' [then a few questions about drugs/weapons/etc. in the car]. (just to clear things up, it's not that my speedometer is "wrong", I also get the speed reading on the phone that I use for navigation. I was definitely not going 74, or anywhere close to 70 even). I also have a dash-cam with my speed... so yeah, definitely 65 [the mpg is crap at 70, and that matters on long trips].

Anyways, I'm guessing that 1) I was moving slower than the rest of the traffic. 2) I was in Illinois with NYS license plate [an odd-ball car], and 3) they found an excuse to pull me over and run my driver's license through the system. And I got a ``warning'' (not a ticket) that says I was going 74 in 70-limit. ...And if they smelled weed or something (? maybe if I was jittery, or didn't follow their instructions ?) they would've had probable cause to search the car for drugs, weapons, etc.

So yeah, you can get pulled over for going 5mph above OR BELOW the limit; or for any other excuse :-/

- Alex; 20220710

Started Longs Peak hike around 5am-ish. Being around Rockies for the last few days helped quite a bit: last time in 2015 this hike took 24-hours. This time around, I was up on top in about 6 hours, and down in about 5, for a total time a bit under 12-hours. This hike is amazing: much more climby than Katahdin (a lot of scrambling and pulling yourself up by your fingers).

It was also a much better day for the hike than in 2015. Then I was one of two hikers (that was it). It was very cold and windy. This time it was windy in parts and cold at times (enough for me to put on a winter hat). But then the sun warmed things up and it was very hot on the way down. There were also hundreds of hikers, with most making it up the mountain (in 2015, most were turning back after the keyhole, due to wind).

All in all, really enjoyed the hike.

- Alex; 20220709

Drove to the Alpine Visitor's center of Rocky Mountain National Park. That's the one that's right on the high-mountain pass at 12k feet elevation. Spent a few hours there---mostly to get aclimated to the altitude.

Then drove to Longs Park trailhead. Found a parking spot this time (last time in 2015, I had to park on the road a bit away---as the road next to the trailhead has a 'no parking' signs). Anyways, there's WiFi at the trailhead---so it's pretty comfy. Can camp (first-come-first-serverd; too much hassle unpacking the tent, etc.; and it is pretty chilly at night at this altitude), or car-camp (which is what I'm doing). Will start the hike tonight around 2am.

- Alex; 20220708

Towards the end of the day decided to drive up Pike's Peak. It seems you need a reservation to drive to the summit---and without a reservation you can only drive 3-miles before the summit. So that's what I did: drive to the parking lot by Devil's Playground trailhead, and hiked up the remaining 3-miles to the summit.

The hike was awesome! I've never hiked Pike's Peak, and I think these last 3-miles are the best: above tree line, rock hopping, etc. There was also a thunderstorm, with lightning and hail and wind, it was scary and fun (well, after the night a few days back, I started to think I'm gonna get hit by lightning on this trip :-/

The summit had some major renovation from about 3-years ago; the old building is gone, replaced with new visitor center, along with new trails, new summit marker, etc., literally everything on the summit is ``new''.

Hiked down (at my own pace), and then saw a park ranger who said that the ``gate-is-closing-at-8pm'', so had to hurry back to the car (my car was the only one in the parking lot at Devil's Playground), and drive-drive-drive. Made it down by 7:52pm... though I doubt they literally 'lock' the gate---they probably require you to get out of the car and just latch it open (as they do in many such places---I once had a locked gate on Mnt.Mansfield in Vermont, and the gate wasn't "locked", just latched close. Anyway, if it was locked, it would've been awesome-cool: as I would've tent-camped on top of Pike's Peak---something they don't let you do these days.

- Alex; 20220707

Towards the end got to Great Sand Dunes National Park. Hiked to the summit of what appeared to be the biggest dune. I think they move, so every year, it's a different dune (?). Anyways, it was awesome.

Then looked for a campsite: the NP campsite was full, the waterfall campsite was full, so the trusty iOverlander app found a campsite that's only 20 miles south of Great Sand Dunes NP: it's a place on a side of a mountain, with amazing scenery. I'm really liking this app.

- Alex; 20220706

Happy b-day! You know who you are...

Meeting my phd advisor for lunch. We haven't physically met in years (due to pandemic, and the fact that he retired and moved to New Mexico). It was awesome---it's very refreshing to see him doing research well into retirement.

At canyonlands met a fellow hiker who suggested using an iOverlander app to find free camping sites. Used that app to find a nice BLM land place a bit north of Santa Fe. Night well spent.

In the evening, lightning struck maybe 200 feet away from car. Then several hits nearby. It was awesome and scary at the same time. Drove a bit downhill after that. Slept in the car---didn't think it was a good idea to use a tent in a potential lightning storm.

- Alex; 20220705

Happy 4th of July!

Since Natural Bridges Monument, got inspired to see the other arches: so decided to sneak in before 6am into Arches National Park. (after 6am, you need that reservation, before 6, they just take payment, and let you in). Anyways, I wasn't the only one with such a crazy idea, and there was a queue of cars to enter arches before the reservation system kicks in.

Hiked to Delicate Arch, and then to Landscape Arch. I think these two are the coolest in the entire park. Didn't spend much time, and headed to Canyonlands, Islands In The Sky.

The goal at Islands In The Sky, was to drive the Shafer Road. So did a quick scenic hop to Mesa Arch (I'm doing arches this weekend), then a bit on the rim walk, and then...

Shafer Road: It is one twisted road that goes from the islands in the sky all the way down to Moab. Anyway, the roads starts out gently enough (and I've seen family in a mini-van going down). Then after a while, the road gets a bit harry---don't think that minivan would make it; hopefully they turned back).

Anyways, after the Shafer trail encountered road-side attractions where you can see dinosaur footprints, etc.

Then into Moab for lunch, and quickly to... that arch that's right next to entrance to Needles. (arches weekend).

At Needles, noticed an open gate to a dirt road to see the Colorado river. I've already done a few dirt roads, what's the harm in another one? Well, this Colorado-river-outlook in Needles road starts out gently, but then gets very weird. Towards the end, it's not a road, but a trail on top of rocks, that somehow a 4wd car just manages to get through.

I'm very surprised at how capable our 4runner is. It's just factory standard. No lift kit, no off-road tires, etc., but it managed to get through that very technical road---where I was thinking the car would flip over, etc., but it just kept on crawling. Now, in all honesty, I have no idea what other 4-runner trims, like trd-off-road package, would offer over the stock 4runner that would make this kinda off-road driving easier....

Anyways, finished up that off-road adventure early enough go do the Confluence trail (or at least half of it). So walked around Needles until sunset. A day well spent.

- Alex; 20220704

After Valley of the Gods, continued up via UT-95. On the way encountered Forest Gump hill (the hill where in the movie Forest Gump is running with a crowd behind him). Anyways, it's a local tourist attraction, so asked someone to click a picture of me running up that hill :-)

Then encountered Natural Bridges National Monument. I didn't know such a thing existed: a less-touristy competitor to Arches National Park. The drive loop has three major "bridges" (arches), and I hiked down to two of them. Pretty scenic, considering there were a total of perhaps 6 cars in the "monument".

Then got to the Canyonlands Maze, and this time I'm approaching it from the south: which is a different dirt road then the one I used before. Anyways, I was the only one at the horseshoe trail---literally didn't see anyone until I got back into a town. Hiked down the canyon, walked-and-walked in the hot-hot sun (in sandals!), clicked pix, and walked back.

Something I didn't notice before, but right next to the registry box, there's another box with binoculars---so can get a closer look if needed.

Definitely regret not wearing boots: sandals are very comfortable, but that sand is stupidly hot, and that hot-sand gets into the sandals, making walking a very bad experience. It was so hot that I drank 4 liters of water during that hike.

After the hike, took the north dirt road; which pretty much terminates at Green River. Found Loves, took shower, had snack, etc.

At Green River gas station, around 2am, when asking attendant ``where can I get something hot to eat'' the gas station attendant just said: ``welcome to America, nobody wants to work anymore!'' Guess that answers my question :-/

- Alex; 20220703

After the grand canyon trip, driving to Canyonlands Maze, via UT-95. It is one awesome road. On the way, encountered something called Valley of the Gods, which is BLM land, so camped there for the night. It was the most amazingly scenic camping night I've ever had. Used my alcohol stove to cook 3 packs of noodles for dinner :-)

- Alex; 20220702

To summarize the last few days: I've been mostly following Route 66 from Illinois to Arizona.

Starting the grand canyon hike. But due to the work schedule (half day is 2pm NY time, which is 12pm arizona time), started the hike later then usual. Also, I didn't sleep the previous night due to wanting to get to the trailhead on time (it's surprisingly difficult to drive from NYC to Grand Canyon and get here within a few hours of planned time).

After starting the hike late, by the time I got to the base of North Rim (cotton wood campground), it was already 6pm, and not wanting to hike in the dark (at least farther away from the car), decided to turn back. So my rim-to-rim-to-rim attempt just turned into rim-to-cottonwood-to-rim :-/

Anyways, because I was passing out from not having slept the previous night, I found a place to nap... the usual spot is on the bench by Phantom ranch, but an even better place is on a bench by Tipoff Point. Next time, I'll plan to nap there on the way down, and on the way up for a few hours.

- Alex; 20220701

Packing for an extended workation (work with a traveling vacation: WFH is awesome, as you can work from a campsite or a hotel).

- Alex; 20220627

Today is the Mnt.Katahdin (in Maine) day. Got to the Baxter State Park entrance at 5:30am, got into the park almost exactly at 6, and started up the Katahdin Stream trail shortly after that.

I haven't done Katahdin in a few years, and I forgot how crawly it is. It's a very fun mountain to climb.

- Alex; 20220626

Hiking up Mnt.Washington (in NH). Doing the usual boot-spur up, then camel trail to the lake, then summit, then lion's head back down. Took almost the entire day, and it was awesome.

Discovered something really cool about the Pinkham Notch: they got a shower. You feed it a $1 in quarters, and you can shower right after the hike.

Next stop: Katahdin

- Alex; 20220625

Wow: Supreme Court overturns Roe v. Wade.

In other news, Suneli/Ian/Liam are flying to India.

In still other news, I'm heading to Mnt.Washington in NH.

- Alex; 20220624

Doing a camping trip to Smith Point County Park. The tent site is just a short-walk from the beach. The camp-site was packed, with spaces setup a bit too close---we had noisy neighbors all night.

The place was full of ticks :-/

- Alex; 20220604

Doing a camping trip near Watkins Glen. The camp-site is just off Seneca lake, within walking distance of Watkins Glen state park.

We got a big Coleman 6-person tent for this trip (we couldn't find our old 4-person tent). It's a bit too cumbersom to setup, but all-in-all, not-too-bad (though I think the 4-person tent would've been a bit easier to deal with).

- Alex; 20220527

Visiting The Steven F. Udvar-Hazy Center in Virginia. That's the place for space shuttle Discovery and SR-71. Awesome place to visit.

- Alex; 20220424

Visiting Jekyll Island, and Driftwood Beach.

- Alex; 20220422

Visiting Daytona Beach.

- Alex; 20220421

Visiting Everglades National Park.

- Alex; 20220419

Visiting Kennedy Space Center.

- Alex; 20220417

Doing a road-trip to Florida.

- Alex; 20220415

Heading home from Virginia.

- Alex; 20220306

Doing a trip to Virginia Beach.

- Alex; 20220304

So here's an idea: the brain is a learning machine that is trying to minimize surprise. A surprised brain burns more energy than a non-surprised brain: so the brain does its best to model the world in such a way that would cause it to be surprised as little as possible. That said, the brain predicts everything: the physical location of every thing in the house for example: if something is moved, the brain is very quick to pick it up: like the salt shaker is on a different shelf, etc. But it doens't have to be the salt shaker, it could be one of other hundred things---so the brain precits the location of the salt-shaker AND those other hundred things, and for the most part, it is not surprised---it's only when the predictions don't match reality is when surprise happens (and brain has to exert energy to learn the new model, or fix environment to conform to the original model; like moving the salt-shaker back to its location).

One way to view Entropy is as an average amount of surprise. So the brain is trying to minimize that: it takes high entropy observations, and uses memory/calculations to turns them into low entropy observations.

- Alex; 20220222

Finished Dear Esther. It came up on "sale" [as free] a few days ago, so just grabbed it. It's a really nice game. Amazing visually, and a nice story---very unique game play. When playing world-of-warcraft, I oftne place more value on just walking around and exploring. Dear Esther is all about just walking around and exploring... there are no enemies, no puzzles to solve, just walk around and explore.

- Alex; 20220218

Finished A Thousand Brains: A New Theory of Intelligence. There are lots of things that I both agree AND disagree on in this book. The modeling aspect of the brain is explained with just-enough-detail to make it interesting, but not enough detail to actually code it... which is OK, since nobody knows how to make the brain stuff work yet.

The assumption about artificial intelligence as being benign, because it doesn't have our primitive brain---and no intentions unless we program those in, that I don't agree with. I think the paperclip scenario is very much real. It may not be paperclips though, but something a bit more subtle: and that's the point, it's not going to be something obvious that we'd notice. This sort of unintended misalignment happens all the time with companies and laws---it's kinda naive to think that it won't happen to AI. Everything that can be misinterpreted, will be misinterpreted in the worst way possible.

Regarding the brain as being a modeler of reality---without any particular desires, that may be true. The desire to "stay alive" and "eat food", etc., might just be coded in the primitive parts of the brain---the "self" may reside in that primitive portion. The higher level brain may not really care for this "self" state. The book essentially makes the case that we're a very primitive reptile brain, with amazingly good world-model capable mechanism. The primitive brain gets hungry/scared/angry, etc., and the higher order brain figures out the smart-solutions to those issues. Without that primitive brain that evolution created, intelligence doesn't really have hunger/fear/anger, etc., So we shouldn't necessarily expect intelligent aliens to have those traits, etc.

I disagree on actively broadcasting our location to the universe. The dark forest scenario from The Three Body trilogy is quite likely... assuming alien live does exist, etc. (with technological evolutions, it is much less risky to destroy any life you find before they have a chance of destroying you).

One really memorable thing from the book was that "knowledge always grows". e.g. once you learn a fact, it stays with you. For example, once Newton learned how gravity works, we never went back... then came relativity... and again, we never went back on our understanding ---yes, folks can still use F=ma, but we know this isn't how stuff "really" works. Anyways, we still don't know how stuff "really" works, but there's definitely a direction to knowledge. This kinda reminded me of the other "direction" in the universe: Entropy. So entropy always increases, and knowledge (and perhaps intelligence) also always increases. There might be occasional reversals, but on the whole, the universe would be getting more entropy AND more knowledge as time goes on. What is this knowledge: models. Models will get more and more correct/accurate/precise. Unless they're destroyed and forced to start new, the models will keep on growing their knowledge indefinitely.

Imagine a model that learns from reality---one that isn't destroyed by accidents, etc., we can pretend that it will eventually learn all there is to learn. So as the universe entropy is increasing, more and more universe's knowledge is very accurately modeled. Anyways, I found that thought process kinda neat.

- Alex; 20220217

Does intelligence imply self-preservation? In other words, if something in intelligent, can it be indifferent if it continues to exist?

This is in the context of AI---will it necessarily require self-preservation (and fight human race for it?) or will it be OK existing with the possibility of being turned off.

What makes humans want to exist? Is it an `intelligent' reason, or is it hard-wired into our brain to resist being shut off?

- Alex; 20220131

For the first time in my life, I'm no longer a student :-/

- Alex; 20220124

Happy New Year!

- Alex; 20220101

Been spending a few days in Virginia Beach, Virginia.

- Alex; 20211230

Been spending a few days in Myrtle Beach, South Carolina

- Alex; 20211228

Been spending a few days in Daytona Beach, Florida

- Alex; 20211226

Been spending a few days in Clearwater, FL.

- Alex; 20211224

Been spending a few days in Savannah, Georgia.

- Alex; 20211221

Driving out on a road-trip to Florida.

- Alex; 20211217

Tried out Windows 11, and after having played with it for a few days, I must say it's the best Windows I've ever used. So much so, that I'm gonna give it a try, instead of wiping it and installing Linux. Best features: WSL, with ability to run Linux GUI apps. So essentially I probably won't miss Linux when using this thing. I do have a bit of a security concern, but it seems (from what I've seen online), the Windows partition is encrypted, and there's no easy way to get at it without a password---so hopefully that means it's as secure as Linux running on the same hardware.

- Alex; 20211215

Got a 4-day weekend: spending it at Virginia Beach.

- Alex; 20211125

Got the flu vaccine at walgreens.

- Alex; 20211124

Happy Birthday!

- Alex; 20211122

Happy Birthday Liam!

- Alex; 20211112

Doing a weekend trip to Brattleboro, Vermont.

- Alex; 20211106

Happy Birthday Ian!

- Alex; 20211105

Finally got around to fixing the TPMS sensor in our car. That tire-pressure light has been very annoying. Here's a short recap of what I've learned about TPMS and wheels in general:

There's a sensor in each wheel. In fact, the valve is part of the sensor. The sensor has a built-in battery that is not user-serviceable (once battery dies, sensor needs to be replaced). Battery life is about 10-years (our car is 2010 model, and battery in one of the sensors just died a few months ago). The sensor broadcasts pressure and its identifier (each sensor has a unique code). The car computer needs to know that code in order to pickup pressure from sensor.

So the step to replacing a dead TPMS sensor is: change it out inside the wheel (more on that in a bit), then reprogram the car computer with the new sensor id. This reprogramming is often done by the dealer (toyota dealer quoted me $100 per wheel to reprogram it; if I change out the sensor myself, or to replace/reprogram TPMS sensors on all wheels: $600). Each sensor is ~$32 on amazon (or over $100 if you buy from dealer)... so 4-wheels ~$130 in parts and $100/wheel in dealer reprogramming--$530 vs just paying the dealer $600 for the whole thing. I'm sure there are cheaper places than the dealer. Anyways, 3rd option: spend $200 on a reprogramming tool (Autel MaxiTPMS TS501 TPMS Relearn Tool Automotive Scan Tool).

Now to replace the sensor: involves removing the tire, or at least part of a tire from the wheel. Never having done that, I didn't think that would be hard---but apparently there's something called ``breaking the bead'' (youtube for it), which is very hard to do without a proper tool. Luckily, there are lots of portable tools for this too: BeadBuster XB-452 Automotive/Car/Truck Tire Bead Breaker Tool

Once the bead is broken, an old 2x4 beam (found in nearby trash) can be used to wedge the tire just enough to expose/reach the old sensor, screw it out, and screw in the new sensor. The just re-inflate the tire, and reprogram the computer using the Autel tool. Done.

All in all, a fun experience. Not sure it saves money though... e.g.: $140 bead breaker tool + $130 for 4-sensors + $200 for Autel tool, same ballpark as the dealer or other shops---but I get to keep 2 tools, and gain valuable knowledge of how to do it :-)

- Alex; 20211031

Just realized that Google's index apparently rolls-off things (or forgets things). In the early 2000s, when you googled for "mona lisa", this website (and linked image) were top search result. Now it doesn't come up unless you search with "site:" attribute, and even then, it finds the wrong url :-/

- Alex; 20211027

Happy b-day to yours truly :-)

- Alex; 20211005

Went to Seven Ponds Orchard to pick apples. Liam walked a lot---more walking than crawling.

- Alex; 20211003

Liam walked across the room without falling.

- Alex; 20210926

Hmm... On the Opportunities and Risks of Foundation Models.

I'm not against the concept of a foundation model---for example, when born, our brain is probably a foundation model, out of which we `specialize' into doing whatever it is we're doing with our lives (e.g. some folks become doctors, others become computer engineers, etc., and yet at the beginning our brains is likely pretty much the same; same capabilities, etc.).

That said, would I think that GPT-3 (or other similar models) are in some sense foundational? Nope. There's a distinct lack of understanding and intelligence.

- Alex; 20210916

Liam walked a good 5+ steps today.

- Alex; 20210915

Been thinking about the nature of the universe lately...

"Spacetime tells matter how to move; matter tells spacetime how to curve." -- John Wheeler

What if it's a 3rd thing that causes both? For example, what if matter is present and space is bent where this 3rd thing is present?

- Alex; 20210908

We found an awesome spot: Little Wolf Beach (campground and picknick grounds).

Plan was to drive to Alexandria Bay; ended up at Watertown NY.

- Alex; 20210905

Had an awesome BBQ by Lake Champlain, by Crown Point.

Staying by Plattsburgh NY.

- Alex; 20210904

Going on a road trip. Staying in Albany NY.

- Alex; 20210903

Tried Quake II RTX. Feels sluggish on a GeForce RTX 2080 :-/

- Alex; 20210824

There's apparently tropical-storm Henri approaching north east. We drove through some heavy rain in NJ---some streets were very flooded.

- Alex; 20210821

Liam took 3-steps :-)

Staying in Clear Spring, MD.

Saw a very clear shooting-star from hotel parking lot.

- Alex; 20210820

Got kids (and myself) tested for covid, and test came back negative. So it must be something else: like a regular cold perhaps.

In other news, changed transmission fluid, front/rear gearbox fluid, transfer box fluid, etc., for first time in 106k miles. It's weird, but the drive feels slightly better after the changes.

- Alex; 20210819

Kids woke up with 104 fever :-/

- Alex; 20210818

Visiting the `New River Gorge Bridge' and driving the Fayette Station Rd back across the river.

- Alex; 20210817

Visiting the `Grandview overlook' of New River Gorge National Park.

- Alex; 20210816

Liam took his first 2-steps.

Continuing with the trip, staying in Beckley, WV.

- Alex; 20210815

Progressing towards WV. Next stop Verona, VA.

- Alex; 20210814

Driving out on a West Virginia road-trip; first stop West Chester, PA.

- Alex; 20210813

RIP

- Alex; 20210808

Used electric scooter (ninebot ES3) to go to manhattan and back. That's about ~9 miles each way [over Williamsburg Bridge]. It was an amazing "ride"; but the scooter battery almost ran out on the way back (started fully charged).

In regular-mode, the top speed is 15km/h, and scooter struggles uphill. There's a sport mode, which ups the speed to 20mk/h and gives it a bit more power uphill... but apparently the scooter burns through batteries pretty quick. Used the speedy mode to get to the destination---but after ralizing that battery was half-dead, stayed in economy mode for the return journey, except for the last mile or so.

- Alex; 20210802

Changed the liftgate support struts on our 4runner. The other day oil just squirted out of it without warning, and that was it for the hatch support. Good things these things are cheap; got a replacement from Amazon for about $30.

- Alex; 20210726

Suneli became a US citizen. Yey!

- Alex; 20210720

Liam got his first tooth.

- Alex; 20210716

Liam said ``MA-MA'' for first time.

Some of Ian's words:
water: watty
straw: stross
straws: strosses
ketchup: dipit
spicy ketchup: picy dipit
sidewalk: wakinger
entrance/outance

- Alex; 20210711

Doing BBQ in Shenandoah National Park.

...and back home in NYC.

- Alex; 20210710

Doing BBQ in New River Gorge National Park, by Sandstone Falls.

Staying in Staunton, VA.

- Alex; 20210709

Staying in Princeton, WV.

- Alex; 20210708

Staying in Wytheville, Virginia.

- Alex; 20210707

Staying in Johnson City, TN

- Alex; 20210706

Staying in Maggie Valley.

- Alex; 20210705

Staying in Pigeon Forge, Tennessee.

- Alex; 20210704

Staying in Bristol, Tennessee.

- Alex; 20210703

Going on mini-road trip. Staying near Allentown PA.

- Alex; 20210701