Joined March 2018
- Tweets 8,880
- Following 0
- Followers 758
- Likes 14,495
429 Photos and videos
17h
đź¤
I am geniously impressed by Rivian and RJ.
Tesla needed over 100,000 H100s running 24/7 inside a custom GPU cluster using the real silicon for bit exact inference during training, years of feedback from their over a million FSD users to improve their models, and over 10 billion supervised miles of data.
And Rivian's going to do it. Without any of this.
They must be geniuses.
22
Steve retweeted
19h
You didn't hear Friedberg and Sacks explain it thoroughly in that very episode?
The cheaters were withholding votes until Bass was locked, then poured most of the remaining manipulated votes into her friend, to block the very popular independent candidate.
With all of the process capture, the only question is how much of the cheating was illegal.
The unarguable part is that the new system is blatantly designed to allow cheating and to be impossible to audit.
2
34
648
19h
I know a lot is going on right now, but isn't @cybertruck Powershare V2H (vehicle to home) meant to arrive OTA any day now?
1
35
19h
I mean the version that works with Powerwall 3 and a Wall Connector (and no additional hardware).
12
Steve retweeted
22h
Some reasons:
1. Software has zero incremental cost for more units. So, sometimes distribution can increase far more than expected. Wider distribution means once rare use cases can become commonplace, sometimes revealing new bugs.
2. Software is often designed to work on wide range of hardware configurations. Even if the spec remains the same, subtleties in new hardware or new combos can break things.
3. Most software is not monolithic, i.e it relies on 3rd party libraries. Those APIs drift over time and they can depreciate features you rely on or break your software.
4. Competition may demand a constant stream of new features delivered by updates, for you to retain market share. This in itself is maintenance, but also creates a long tail of QA and can introduce new bugs that have to be fixed.
Also, hardware does have maintenance, but it manifests differently. Something may be deemed an acceptable bug in hardware if it is cheaper to handle through warranty claims than manufacturing or design improvements (and if it won't burn too much reputation).
Or the maintenance is just passed on to the owner.
Or the improvements are batched for a next rev, e.g. car model years.
Etc.
1
1
24
Jun 12
Finally watched Sin City 2: A Dame to Kill For. It's much better than the 6.5 IMDb it has.
As a fan of graphic novels, I like the interpretation and cinematography. The first Sin City movie mostly worked too, but turned me off in some way that I can't recall.
Goddamn, Eva Green and Powers Boothe were insanely good villains. I liked all of the cast, though the Mickey Rourke facial prosthetic was hard to get past.
Was also fun to see a few of my favorite actors from before they got big, notably Julia Garner.
The 2014 SFX hold up well. Obviously, you have to set in with a little extra suspension of disbelief.
81
Jun 12
It seems like there is a weird scenario possible under Schwab IPO rules, where you could get an allocation of $SPCX via COTP and then optionally buy it or not(!) on Monday at the IPO price ($135) while already knowing the open market price for it.
I don't want to describe the mechanisms publicly. But can anyone confirm or deny that this is possible?
1
1
1,171
Jun 11
The affirmation window for $SPCX IPO at Schwab is now open.
Note that the "Affirm Now" option looks final for your originally requested amount, but after clicking that, you get an option to modify (reduce) your requested quantity if you wish, before confirming.
Act fast, you only get tonight to lock it in. Then you find out what amount you get to buy in the morning.
2
2
670
Steve retweeted
Jun 11
Best practices for making crypto recoverable upon death
option 1: Leave keygen words in a safety deposit box or safe that the family will be able to access.
Option 2: Set up an M of N multisig, e.g. that can be accessed by a beneficiary and your lawyer, or by two of the several beneficiaries.
There are also custodial options, but it's foolish to use a custodian.
1
256
Steve retweeted
Jun 11
Karpathy had an early interest in generative code. Maybe he saw that Anthropic was approaching meaningful RSI. So the window was closing for influencing the outcome more directly.
Elon's statement about his evil detector calming down may have led K to believe that meaningful altruistic influence at Anthropic to was still possible.
1
1
149
Jun 11
A new Apple route sharing API this fall may enable coordination with FSD. Will Tesla then allow use of CarPlay navigation?
And where is the Tesla nav routing overhaul and upgrade that build decompiles have hinted at?
Meanwhile, Tesla nav routing (ironically) remains a noteworthy weak point to the FSD experience and a major source of frustration and interventions.
Jun 11
Apple this week announced a new Maps feature that could be the missing piece in Tesla's rumored efforts to officially bring CarPlay to its vehicles. Learn more: notateslaapp.com/news/4278/c…
1
201
Steve retweeted
Jun 11
The metric I keep coming back to for SpaceX is $/Mbps to orbit
Starlink exists because Falcon 9 dropped bandwidth deployment costs ~10x to ~$6.55/Mbps. That’s about to drop again to just $0.30/Mbps because of Starship.
A business that is doubling users annually with a 63% adjusted EBITDA margin is about to cut their biggest cost by 95%… It really seems like people don't understand the implications of this.
The math assumes a reusable Falcon 9 launch is 17 tonnes at $1,000/kg and 2,600 Gbps per launch. Starship is targeting 100 tonnes at under $185/kg and 61,000 Gbps per launch. That's $17M for 2,600 Gbps ($6.55/Mbps) verse $18.5M for 61,000 Gbps ($0.30/Mbps).
Starship's additional volume allows for larger satellites, enabling simultaneous gains on multiple cost curves. The math suggests V3 satellites are ~600 Mbps/kg vs ~150 Mbps/kg from V2 mini.
Combining the 4x improvement on satellite bandwidth density with a 5x improvement in launch gets you the 20x improvement to 30 cents per Mbps to orbit.
These are fairly conservative assumptions because launch probably comes in even lower as Starship ramps, and satellite improvements probably keep coming. At $0.10 / Mbps, $1 billion spend on launch represents 10,000 Tbps or about 15x the bandwidth of Starlink's constellation today.
$1B is 90 days of operating income for Starlink... at it's current scale...
Yeah, I really don't think people are getting this. Starlink is the internet now.
195
404
4,063
360,818
Jun 10
Many countries WERE substantially enhanced by immigrants. But that was before unlimited handouts, open borders, and insane political capture that turned officials into virtue signaling racists.
Every developed country needs immigration reform to combat damage from years of vote importing and cultural destruction by leftists.
Immigrants can come if:
1. The country is not in distress and we have control of our borders
2. They learn the language, the culture, and the laws
3. Have significant basis for being self-sufficient
...just like every other country!
Non-citizens should be immediately deported if they:
A. Participate in political movements without registering as foreign agents
B. Commit crimes (violence, vote fraud, intimidation)
C. Become a chronic drain on welfare
...just like every other country!
Further:
- end birth right citizenship; it's unique in the world and it wasn't meant for this
- remittances should be taxed heavily
- asylum should be paused
1
70
Jun 10
Exceptions should be made for high performers in important fields, but the hiring company should have to pay a high fee and an ongoing fee, enough that qualified domestic hires are not disadvantaged.
23
Jun 8
I miss Cane's chicken, but I haven't been back since the insane shrinkflation. There are decent alternatives that aren't a rip-off.
Reposting Cane's sauce recipe again because it's the best. Very easy to make.
1 cup mayo
1 tbs ketchup
1 tsp garlic powder
Black pepper to taste
Worcestershire sauce to taste
Stir to mix well
Serves 4 to 6, but easy to reduce
22
Jun 8
Why do so many people day trade their taxable brokerage accounts, churn up an insane amount of short term gains, and then let their tax sheltered 401ks just rot for in some shitty bank-sponsored mutual fund?
1
3
29
Jun 8
Our new and improved DMV system is funny.
You value your time, so you make an appointment. No time slots for a month? Okay. Finally the day arrives, you show up 30m early and check in. Then... they just put you in the same line as everyone else!
Appointment time comes and goes. Finally your number is called and then you find yourself standing at a window next to a guy who arrived after you and just did a walk-in with no appointment.
🤔
41
Jun 8
Nasty
Jun 8
Vinod’s trigger happy impulse to can founders has dredged up what happened to me. I’ve never written about this before, so here goes.
1990s dot com boom. I had solely founded Websense, a profitable (from the get go) software Internet company in 1994. We didn’t need VC money, but everyone and their dog food company was getting VC money so I decided to get investment as a defensive measure in case one of our competitors also got a large pile of cash.
Very long story short, I ended up with Morgan Stanley Venture Capital out of New York leading our series A in late 1998, odd considering we were in San Diego. Anyway, post investment, the deal partner made it clear he wanted me out. Long series of maneuverings, but I got fired from my own company, and then had to battle against Morgan Stanley who was angling to do a cram down round to extract even more equity from me.
I found a white knight, ex-corporate lawyer, sharp business person. I gave him some of my equity to go to bat for me. Next telephonic Board meeting, before Morgan Stanley could propose their cram down, I resigned and put my white knight in my place on the Board. You could hear a pin drop. The credible threat was made, Morgan Stanley played it more or less straight from then on.
Yes, we were a young inexperienced team, but we had built the company profitably to $10M in yearly revenues. Their new management wasn’t magically better. They actually missed a bunch of strategic layups. At least they finally took the company public - one week before the NASDAQ freefall in late March 2000. Sigh. After I had sold all my stock, they continued doing nothing with the company so much that it got taken private by a PE firm. That’s a whole ’nother story.
I did fine financially. But there is no doubt had I gone with a founder friendly VC, we would have executed better, and had I not gone with any VC, but kept growing organically, I would be a far richer person today (and probably had a stress heart attack to go with it).
Yes, there are cases where a VC should replace the founder CEO, like if they lie to the Board, do major deals without Board approval, or are seriously going off the rails in their private life. But absent that, realize that a replacement CEO usually isn’t founder material themselves, else they would be running their own startup.
And spare me Vinod’s “I do what’s right for the company” BS. That’s a thin cover for “doing what’s right for my VC firm”. Never waste a crisis, and somehow deal makers always manage to extract an extra pound of flesh in any restructuring they’re involved in.
I’m not in the same league as other founders who got shitcanned like @PalmerLuckey, but needless to say, I can relate 100%.
Yes, your founder is young and inexperienced. But they ate the broken glass and built what you’re so wanting to invest in. Don’t fall for the “grass is greener” fallacy and think you can slot in a better CEO.
1
48
Jun 7
Bold
Set is ready @ElonMusk
We built it 25min from downtown Austin and can shoot anytime in the next 7 days on 1h notice.
Humanity is on the verge of becoming a multi-planet species and spacefaring civilization.
My goal with this interview is to help people viscerally feel what that future is going to look like and get everyone excited to help build it.
2
131