Tesla fsd

 

🚗 

A Long Drive with Tesla’s Mind: My 10-Hour Journey from El Paso to San Antonio

When I left El Paso around 4 p.m., I wasn’t sure whether I trusted my car.

The plan was simple: drive to McDonald’s myself, get food for the family, and then let Tesla’s Full Self-Driving (FSD) take over for the long haul. But simple plans feel different when you’re about to hand control of two tons of metal to a neural network.

First Miles: Apprehension

The city stretch was a test of nerves. I kept waiting for FSD to misread the chaos—cones, medians, impatient drivers cutting in, merge lanes that appear and vanish. But it didn’t. It glided through the noise of El Paso’s evening traffic like a cautious but confident human, obeying lanes and reacting fluidly to other cars. Each correct decision chipped away at my anxiety.

Out of the City: The Machine Finds Its Rhythm

Once we hit open highway, the system seemed to breathe easier. It sped up smoothly, kept a steady distance, and even overtook lumbering trucks with the same subtle confidence you’d expect from a seasoned driver.

Then came nightfall. I thought the darkness would confuse it—headlights, reflections, shadows—but it handled the curves and the glow of oncoming traffic like it could see in the dark. FSD moved aside for faster cars, keeping the flow natural. Somewhere along that stretch, I realized I’d stopped second-guessing it. My hands hovered less. My eyes roamed more. The car was driving, and I was supervising.

Charging: The Modern Pit Stop

At the first Supercharger station, FSD hesitated—it stopped in the middle of the lane, unsure where to park. A quick scroll-wheel selection fixed it, and it reversed perfectly into a stall.

That small handoff captured the current state of AI driving: brilliant at motion, still learning intention.

Charging was a breeze. In 29 minutes, the car juiced up enough for the next leg while my kids and I grabbed food and stretched. We repeated the ritual at three more stations. The costs were light—about $10–20 each stop, roughly $200 total for the round trip. A gas car would have burned far more.

Arrival and Reflection

From El Paso to San Antonio—and onward to Webster, near Houston, for Great Wolf Lodge—FSD carried us the whole way.

Ten hours behind the wheel, yet I wasn’t tired. The mental load of driving had dissolved. I didn’t talk to the car; it didn’t talk to me. Trust grew silently, through performance.

By the end, it wasn’t about technology anymore. It was about coexistence: when to let go, when to watch, and how human trust evolves one correct decision at a time.

What the Drive Taught Me

FSD today is both extraordinary and ordinary—it doesn’t feel futuristic once it works; it feels natural. The only time I felt “AI” was when it hesitated at the charger. The rest was a smooth dialogue between human patience and machine precision.

If I learned one thing, it’s that the future won’t arrive with a jolt. It’ll slide in quietly, mile by mile, until one day we realize we’ve been sharing control all along.

 https://open.substack.com/pub/paulkrugman/p/china-has-overtaken-america?utm_campaign=post&utm_medium=email

An article by Paul Krugman today on substack.

From one end of the political spectrum this is usual but I decided to ponder the reverse. Most articles state Trump and the right have got it wrong and the left have it right. No space of compromise.

What if the truth lies somewhere in the middle, as it normally does. In math this is the mean, the median and the mode. Take the most extreme views from the right, and then the left and find what the middle ground is.

Also he talks about the assault on knowledge which brings me to what the truth is. One thing is for sure. new knowledge tends to destroy old power structures, and these structures would do everything to prevent dissemination of the new knowledge. 

Most people do not care about the truth they only care about their truth. What helps them retain power and dominance.

Palindromic numbers

 I just discovered palindromic numbers while browsing TikTok , really amazing to see how they work . 

Thinking of immigration

 Reading opinions on both sides of the recent 100k fee for H1B workers got me thinking about benefits / costs of immigration to any society . 

I couldn’t Marshall my thoughts together just by pondering these opinions which were more of if these immigrants were absent we would take their jobs versus immigrants might be creating these jobs , so I decided to ask chat gpt to compare 2 fictional countries :

Here is the output :

Here’s a clean, apples-to-apples thought experiment that shows what opposite immigration policies can do to jobs, output, overall activity, and social cohesion.

Setup (two hypothetical countries)

• Same starting point (Year 0): population 50M, labor force 25M, unemployment 5%, GDP $2.0T, trend growth 2%/yr.
• Production: Cobb-Douglas (constant returns): Y = A K^{0.35} L^{0.65}.
• Policies (10-year horizon)
  
• Openland (welcoming): net immigration = 0.8% of pop/yr (≈400k/yr). 75% are working-age; 70% of those join the labor force. Modest diversity/innovation bump to productivity of +0.1 pp/yr once absorbed.
• Closedland (restrictive): zero net immigration. Aging trims labor force –0.3%/yr. No extra TFP bump.
• 
  

Intuition: In the short run, more labor nudges wages down a touch and raises GDP; with time, capital and productivity adjust so wages recover and GDP per capita depends on whether productivity benefits arrive. In restrictive settings, fewer workers can lift wages briefly but slow total activity and strain public finances as dependency rises.

Short-run (0–3 years): frictions dominate

Jobs & wages

• Openland: Labor supply rises; with capital fixed initially, sectors exposed to new competition can see a small wage dip (rule-of-thumb: ~–0.2% wage impact for a +0.6% labor shock/yr), and unemployment can tick up ~0.1–0.3 pp while matching/credentialing occurs. Displacement is usually small and temporary; complements (managers, engineers, owners) often see gains.
• Closedland: Tighter labor supply nudges wages up a bit in shortage sectors; unemployment can dip below 5%. But firms face hiring constraints → postponed projects, longer wait times, and more inflation pressure.

Output & activity

• Openland: With K fixed, extra labor raises Y roughly by (1-\alpha)\Delta L ≈ +0.4%/yr above trend. More consumption, more services used; housing and infrastructure face pressure if supply is tight.
• Closedland: Shrinking labor force reduces near-term output growth by ≈ –0.2%/yr vs trend. Some substitution to automation; service availability tightens (childcare, eldercare, hospitality).

Social cohesion

• Openland: Big inflows stress housing, schools, and identity if integration lags → higher short-run friction unless governments expand supply (housing/transport), speed up credential recognition, and invest in language/civics.
• Closedland: Less cultural churn; fewer near-term strains. But labor shortages can widen inter-regional divides (booming, high-rent cities vs stagnating towns), and fiscal pressures build with aging.

So any policies in this regard have to consider this factors . Do not based immigration on just a moral appeal, look to the economic effects.

Long-run (10 years): capital & skills adjust

Using the assumptions above (incl. a modest +0.1 pp/yr productivity bump in Openland), here’s a compact scoreboard:

Started listening to the book "The Singularity is Nearer", simply amazing ideas that Ray Kurzweil, opened with. It is impressive to see thinkers like this and OpenAI creators, ponder on the future of humanity as a whole. Not just for themselves, their immediate families or tribe but for humanity as a whole.

As per this piece of writing, for context, globally, we are already inextricably attached to machines, especially in the form of cellular phones, whose penetration is very high, though not 100%. I think this is part of the singularity, closer man to machine.

I assume thinkers like Ray Kurzweil, imagine a singularity as silicon-carbon based life, where artificial technology is embedded in our biology and we can seamlessly harness its powers. So to make a call I think of my mom and biology tells her brain that I want to communicate with her.

Though the current exciting events around mapping the human brain to silicon and replicating human like responses i.e artificial intelligence, make machines more useful for us. I posit that increasing human computational power via machines is not the needed advance towards the singularity but improving our communication would be the next stage of our evolution.

Many schools of thought exist as to our evolutionary history and what makes homo sapiens different from our common ancestor with chimpanzees. Undoubtedly, advanced communication, using a wider vocabulary (human language), understanding wider varieties of facial expressions, and body postures, has helped advance us.

With this in mind, and supported by the observation, that computer technology has advanced not only when computational power increased but when computers were applied to communication, ala, netscape navigator, facebook, whatsapp etc. Technologies that enhance our ability to communicate would be key to advancing towards the singularity.

While this advance occurs, thinkers have to be cautious of the adverse effects of enhanced communications, destructive memes spread faster. It becomes easier to intimidate and bully people with always on access, if silicon were in our brains how do we shut off communication, without the ability to distance ourselves. Guardrails would be needed before such technology becomes mainstream.

So, AI tools that improve our communication would take us to the next level. Imagine chatting with a AI about a specific topic, and the AI develops themes about the topic, then if another person has a question and you are not available. You can talk to the AI. Better still people interact with an AI over a long period of time and AI records responses and back feeds these responses to itself and can actually replicate the thoughts, emotions, motivations or these people. In a way we would be living forever like this.

More fantastically, would be when genetic modification can be done to gametes and the offspring produced from this can create carbon based semiconductors and have brain signals directly produce radiowaves which can be captured by other wireless devices and transmitted onwards.

 

Applied to OpenAI grove program today.

Just finished reading the book "The Optimist, Sam Altman, OpenAI, and the Race to Invent the Future" by Keach Hagey. It was eye opening.

I had underestimated the value of networking to refine ideas in technological progress and the book highlighted this salient factor. I thought that just plodding through technical details in the hope of solving a problem is all that was required for technological breakthrough.

The technology might be there but the market not, the market might be there but the technology not. Find the balance and mix is key.

I hope they accept him to connect with thought leaders.

As to my Survey bot, currently I am rolling a simple interface that I want to test among some Instructors. Hopefully would send out the links today. 

 woke up early today

decided to do some work

ran yesterday and hopefully would run today