Todd Lefor
1.2K posts

Todd Lefor
@toddlefor
Sarah and Joshua's daddy. Software. Artificial Intellegence. Deep Learning.
เข้าร่วม Ağustos 2010
657 กำลังติดตาม131 ผู้ติดตาม

I believe Starlink will become the largest subscription business in the world and it’s only a matter of time.
This is where the top players stand right now:
• Starlink: ~11.8M subscribers
• Netflix: 325M subscribers
• Spotify: 293M subscribers
• Amazon Prime: 200M subscribers
• Disney+: 132M subscribers
Right now Starlink is still the smallest… but the opportunity is MASSIVE.
There are still over 2.2 billion people on Earth without internet access, roughly 26% of the global population. That’s more than the entire population of the Americas combined.
Traditional infrastructure can’t reach the remote, rural, and underserved regions where these people live. But Starlink’s satellite constellation is built exactly for it with global, high-speed internet anywhere on the planet beamed down from space.
As terminal prices continue to drop, coverage expands, and demand for reliable connectivity explodes in emerging markets, the growth potential is enormous. I’m talking hundreds of millions of future subscribers…
When I say things like this, many people laugh, but I’ve always been laughed at till it becomes true. Starlink is the future of global connectivity.
English

After experiencing life with a Self-Driving car daily, it seriously changes your life in so many ways.
This comes from someone who loves to drive. Heck, I bought possibly the best-handling Tesla there is: the Model 3 Performance.
But 90% of the time I’m either stuck in traffic or commuting somewhere that doesn’t require me to be so engaged. Don’t get me wrong, it’s amazing to have the choice to drive manually, it’s one of the best driving cars out there, but most of the time, sitting back and having my car drive me around so smoothly is just incredible.
It’s something you need to feel for yourself to truly understand.
English

🤯 Nvidia has solved autonomy. 👏
Shay Boloor@StockSavvyShay
$NVDA CEO Jensen Huang just announced Alpamayo which he calls the world’s first thinking and reasoning model built for autonomous vehicles. By open sourcing the Alpamayo stack, Nvidia is pushing self driving forward as a category after years of work by thousands of engineers.
Español

@elonmusk
What do these people have in common? They all explain FSD.
I wonder if my Tesla thinks about me too…
a.co/d/60C0Hip

English

@elonmusk
Learn more about autonomy in Optimus and Tesla FSD
I wonder if my Tesla thinks about me too.
a.co/d/fmXC7F4

English

amazon.com/wonder-Stella-…
With FSD engaged, sudden evasive maneuvers can prompt an instinctive human override. That split second can result in neither the driver nor FSD executing the optimal response. What does Tesla think about this human FSD conflict?
@elonmusk
English

With FSD engaged, sudden evasive maneuvers can prompt an instinctive human override. That split second can result in neither the driver nor FSD executing the optimal response. What does Tesla think about this human FSD conflict?
@elonmusk
English

@elonmusk The trolly problem is no longer a problem when autonomous systems are everywhere… the only winning move is not to play… you avoid near misses by adding FSD to all automobiles…
English

@elonmusk We need to look closely at what parts of humanity we want to instill into AI
English

Generated with @grok imagine
FSD Demystified with humor and historical examples and background. Neural nets as AI minions…
a.co/d/35SI4zx
English

If the pickup also had Full Self-Driving, it would not have attempted to pass the semi at that moment, and the dangerous situation would not have occurred.
That’s how you solve the Trolly problem. You prevent the dangerous conditions from manifesting in the first place.
@elonmusk
English


@lexfridman With FSD engaged, sudden evasive maneuvers can prompt an instinctive human override. That split second can result in neither the driver nor FSD executing the optimal response. What does Tesla think about this human FSD conflict?
English

Doing a long, super-technical podcast on the state-of-the-art in AI. Let me know if you have question, topic suggestions. Everything from details of LLM training pipeline & architectures, to coding, robotics, scaling, compute, business, geopolitics, etc.
Besides topics & questions... add papers, blogs, posts, rants, perspectives that you'd like to see covered.
English

When Tesla finally hands me an Optimus, I’m going to train it everything I cook. I’m going to teach it how my family eats, how real meals are cooked, how timing, taste, love and care is the most important ingredient in making a delicious dish.
And once Optimus learns it right inside my home, anyone in the world with a Tesla Bot will be able to get that same skill through a simple software update. Just from an update, suddenly your humanoid robot knows how to cook the same exact meals.
This is what people still don’t fully understand about Tesla, which is shared intelligence at scale.
The same way Tesla FSD gets better bc millions of cars are learning together, Optimus will get better bc real people will be teaching it real life.
This is going to be the best product ever created, by far. Optimus is going to change everything… I can see it clearly now.

English

@elonmusk
With FSD engaged, sudden evasive maneuvers can prompt an instinctive human override. That split second can result in neither the driver nor FSD executing the optimal response. What does Tesla think about this human FSD conflict?
English

@TheCaptainEli With FSD engaged, sudden evasive maneuvers can prompt an instinctive human override. That split second can result in neither the driver nor FSD executing the optimal response. What does you think about this human FSD conflict?
English

In his 2019 interview with Lex Fridman, Elon Musk brilliantly foresaw that Tesla’s autonomy would soon surpass human driving, making interventions riskier. He used the perfect elevator analogy: “It used to be that there were elevator operators… now nobody wants an elevator operator, because the automated elevator is much safer. In fact, it would be quite dangerous to have someone with a lever.”
Fast-forward to 2026, and the latest Tesla safety data proves Elon right once again.
With Tesla Self Driving (Supervised) engaged, vehicles record roughly one crash every 6–7 million miles—about 9× safer than the U.S. average (~700k miles) and dramatically better than driving without it (~1–2M miles per crash).
As usual, Elon’s visionary insight is spot-on: advanced autonomy is already delivering exponentially superior safety, transforming roads and saving lives.
The future he dreamed of is here!
English

Since using FSD from v12 to present v14.2.2.2, I’ve encountered countless construction zones. Not once has FSD attempted to enter a closed work area or become trapped in a situation like this.
Waymo has a far richer sensor suite than Tesla, but sensors alone don’t guarantee better behavior. Without strong intelligence behind them, sensor fusion offers limited value. In the end, smart decision-making AI matters more than the complexity of the sensors themselves.
No Safe Words@Cyber_Trailer
Phoenix, AZ (JAN. 02. 2025)- Men at work. Waymo enters closed work area to wander around tree cutting crew boom lift and wood chipper. #SunsetProject Video credit - ixchell on TT
English



