Tyler Wasniowski

@tslaming This would be hard complexity to manage (false-positives could degrade overall performance), will be interesting to see if they can pull this off without degrading non-emergency performance.

@zalberico @jawwwn_ @PalmerLuckey @gdb @elonmusk This is too simple. Elon is in the weeds on choke points. His time is limited. Most teams operate autonomously for months with little feedback from Elon. Elon has said before that the best employees don’t need to be managed.

Anduril Founder @PalmerLuckey: You MUST trust your team. “I met with two OpenAI executives at the White House.” “@gdb said, ‘How are things going on [redacted] project that our teams are working on?’” “I said, ‘I honestly couldn’t tell you. I haven’t heard anything about it in a month. I just directed that team on what to do. I assume it’s going well.’” “He said, ‘Yeah I have no idea either. I assume it must be going well.’” “We’re like, ‘Yeah—it must be going well.’” “Maybe it’s not, but if it is, this is a great story.” Via @WestPoint_USMA

@SettySLC @TaxEveryone Performance incentives became more popular for them. The incentives help align the executives with the shareholders. More value is created this way. Since many employees receive stock, this shift is actually beneficial for everybody involved. Capitalism is full of win-win-wins.

@TaxEveryone Then how do you explain the explosion of C suite pay at a disproportionate rate compared to normal employees?

There should be an automatic community note every time this graph is posted

@RealDanODowd @28delayslater @RealDawnProject Where do you live that the sky in the video is a “clear day”

@28delayslater @RealDawnProject There may well be edge cases where FSD still fails, but in the basic scenario of approaching the stopped school bus on a straight road on a clear day, FSD 14 passed all 50 test runs!

BREAKING: Tesla self-driving finally passed @RealDawnProject's school bus test! On 50 out of 50 test runs, FSD 14.3.2 correctly stopped for the stopped school bus. After three years of campaigning, New York Times ads and Super Bowl commercials, Tesla has finally done the right thing and fixed this egregious safety defect. However, HW3 Teslas running FSD v 12.6.4 will still illegally blow past the stopped school bus and run down a child crossing the road. @ElonMusk, if the defects can't be fixed on the current hardware, FSD should be disabled until those vehicles have been upgraded. @aelluswamy, your next task is to stop FSD 14 from speeding in school zones.

@CozendeyMath @VictorTaelin I agree, you can get pretty close to all practical functions heuristically.

@VictorTaelin I don't speak german (is it really impossible though? For every every input of function X that maps to Y function Z should return Y too? If not possible, can't static analysis get you at least to some useful point heuristically?)

apparently there are 409 calls to U32/to_u32 in Bend's stdlib this is a function that converts u32 to... u32. and that, kids, is what happens when you leave AI unsupervised never again. cleanup in progress

@ashiningcity @AviFelman Exactly

@AviFelman no, its the opposite. FIRE is an abundance mindset. you don't need to be stuck in a career you don't like because you are not in financial danger. you will be ok. riches and opportunity are abundant.

FIRE is inherently a scarcity mindset, which doesn’t make sense. We live in an age of abundance for those that want to have it. If you truly want to succeed in life — spend the money. You’ll find that you go further. You’ll be pushed to work harder, to dream bigger.

It’s interesting how one-sided this discourse seems to be. In cases where you have more salary, you can take lavish vacations drive nice cars, etc. And plan to achieve other goals after work (building a business, competing in a sport, building a community, etc.). This is not a scarcity mindset. This is a mindset that we have enough abundance, such that after several years of working, you can live your life as you see fit without taking the highest value offer for your time. The guy in the original post is just a bad example who is over-saving. They are not representative of people who actually earn enough to live a lavish lifestyle, and retire early.

@TeslaUltraBull @TylerHardt It took decades of research to arrive at a complex EUV machine. A simpler one is probably possible but will take at least years for even the best engineers to arrive at. Raptor is comparitively much simpler and took a decade plus to arrive at the ‘simple version’.

@TylerHardt Physic is the law, everything else just recommendation. There is more than one way to build EUV, probably more efficient and more advanced way than ASML.

I notice you failed to mention how $TSLA is going to do any of this without buying an EUV Lithography Machine from $ASML considering they are sold out through 2028? Seems like an important detail. $TSLAQ

@aelluswamy LFG!!! This is a huge moment! Congratulations and thank you to the Tesla team! ❤️

Robotaxi rides without any safety monitors are now publicly available in Austin. Starting with a few unsupervised vehicles mixed in with the broader robotaxi fleet with safety monitors, and the ratio will increase over time.

@farzyness Money will still be relevant. There will still be jobs like entertainers/athletes earning lots of money. Others will have the option to work in service to those rich entertainers possibly with things like massages, tour guides, or other jobs where you might prefer human vs. robot

If there's not need to save money, what will we use to decide who gets what piece of land?

The internet unlocked opportunity if only you were intelligent. AI is unlocking opportunity if only you have agency. One of those is more easily learned.

@Doublelift1 @elonmusk @grok @LeagueOfLegends This is not the matchup I expected, but I’m here for it! Let’s go DL!! 😁

@elonmusk @grok @LeagueOfLegends if I lose to Grok I'll shave my head, no shot this would happen

Let’s see if @Grok 5 can beat the best human team @LeagueOfLegends in 2026 with these important constraints: 1. Can only look at the monitor with a camera, seeing no more than what a person with 20/20 vision would see. 2. Reaction latency and click rate no faster than human. Join @xAI if you are interested in solving this element of AGI. Note, Grok 5 is designed to be able to play any game just by reading the instructions and experimenting.

@karpathy Any measurement would need to take into account the shifting goalposts of slop. As better quality is easier to produce, the bar for slop is raised.

Has anyone encountered a good definition of “slop”. In a quantitative, measurable sense. My brain has an intuitive “slop index” I can ~reliably estimate, but I’m not sure how to define it. I have some bad ideas that involve the use of LLM miniseries and thinking token budgets.

@karpathy Man wait until you get v14, it’s crazy smart

I took delivery of a beautiful new shiny HW4 Tesla Model X today, so I immediately took it out for an FSD test drive, a bit like I used to do almost daily for 5 years. Basically... I'm amazed - it drives really, really well, smooth, confident, noticeably better than what I'm used to on HW3 (my previous car) and eons ahead of the version I remember driving up highway 280 on my first day at Tesla ~9 years ago, where I had to intervene every time the road mildly curved or sloped. (note this is v13, my car hasn't been offered the latest v14 yet) On the highway, I felt like a passenger in some super high tech Maglev train pod - the car is locked in the center of the lane while I'm looking out from Model X's higher vantage point and its panoramic front window, listening to the (incredible) sound system, or chatting with Grok. On city streets, the car casually handled a number of tricky scenarios that I remember losing sleep over just a few years ago. It negotiated incoming cars in tight lanes, it gracefully went around construction and temporarily in-lane stationary cars, it correctly timed tricky left turns with incoming traffic from both sides, it gracefully gave way to the car that went out of order in the 4-way stop sign, it found a way to squeeze into a bumper to bumper traffic to make its turn, it overtook the bus that was loading passengers but still stopped for the stop sign that was blocked by the bus, and at the end of the route it circled around a parking lot, found a spot and... parked. Basically a flawless drive. For context, I'm used to going out for a brief test drive around the neighborhood to return with 20 clips of things that could be improved. It's new for me to do just that and exactly like I used to, but come back with nothing. Perfect drive, no notes. I expect there's still more work for the team in the long march of 9s, but it's just so cool to see that we're beyond finding issues on any individual ~1 hour drive around the neighborhood, you actually have to go to the fleet and mine them. Back then, I processed the incredible promise of vehicle autonomy at scale (in the fully scaleable, vision only, end-to-end Tesla way) only intellectually, but now it is possible to feel it intuitively too if you just go out for a drive. Wait, of course surround video stream at 60Hz processed by a fully dedicated "driving brain" neural net will work, and it will be so much better and safer than a human driver. Did anyone else think otherwise? I also watched @aelluswamy 's new ICCV25 talk last week (x.com/aelluswamy/sta…) that hints at some of the recent under the hood technical components driving this progress. Sensor streams (videos, maps, kinematics, audio, ...) over long contexts (e.g. ~30 seconds) go into a big neural net, steering/acceleration comes out, optionally with visualization auxiliary data. This is the dream of the complete Software 1.0 -> Software 2.0 re-write that scales fully with data streaming from millions of cars in the fleet and the compute capacity of your chip, not some engineer's clever new DoubleParkedCarHandler C++ abstraction with undefined test-time characteristics of memory and runtime. There's a lot more hints in the video on where things are going with the emerging "robotics+AI at scale stack". World reconstructors, world simulators "dreaming" dynamics, RL, all of these components general, foundational, neural net based, how the car is really just one kind of robot... are people getting this yet? Huge congrats to the team - you're building magic objects of the future, you rock! And I love my car <3.

@nearlydaniel I think the fastest will be speech in photo/gif/video/custom-GUI/text out.

Interesting asymmetry in AI interactions: You speak faster than you can type But you read faster than you can listen So pure text is slow, and pure voice is slow “Speak and read” seems like the fastest 2 way bitrate we have until brain implants

@BenjaminDEKR The public knows/thinks Bitcoin is secure. Meanwhile there have been many businesses that use encryption with data leaks due to data mismanagement. The analogy is helpful to the public to convey that the intention is extremely secure messaging.

Okay no shade, but genuinely why does he keep comparing messaging encryption to Bitcoin? It's such a strange, not great analogy

@VictorTaelin @elonmusk @kimmonismus @karpathy No shot. Unless you give it like a whole year which is longer than it took for Karpathy.

@elonmusk @kimmonismus @karpathy so grok-5 will be able to build nanogpt from scratch?

At this point, it would be important to clarify which definition of AGI Elon Musk is using.

@NVIDIAGeForce GeForce Day

🟢 GEFORCE DAY IS BACK 🟢 To celebrate, we're giving away TWO GeForce RTX 5080 Founders Edition GPUs, signed by NVIDIA CEO Jensen Huang. Want one? Comment "GeForce Day" for a chance to WIN & stay tuned for more!

@amilajack 🚀Nice job Amila! But wouldn’t expect anything less from you/Palette. I know you’re just getting started! 🚀

In just one month at Palette we: 1. built out SSO support 2. rewrote our core profile aggregation service in rust for significant perf wins 3. onboarded 3 amazing enterprise customers Couldn't be more excited for what's next!