Amrit__

[Amrit__ @Amrit__M]

@skyler_chan_ Lunar Citizen

[Amrit__ @Amrit__M]

That distinction actually makes sense — dev tool costs are one-time to build a skill, not recurring per task. Curious about the business model though: is the revenue in developing skills once for clients, or in the compounding skill library over time? That’s the part that decides if it scales like a real product or ends up being closer to a services/consulting play.

[Amrit__ @Amrit__M]

@Migma_AI 👋

[Amrit__ @Amrit__M]

@skyler_chan_ Looking forward to it!

[Amrit__ @Amrit__M]

Agree with the post-PMF distinction - hardware stops being the bottleneck once a robot’s actually deployed and working. But I still lean toward in-house software when you want real control and cost efficiency long-term. Companies that own their software, data pipelines, and code end up able to reinvent and expand into new products in a way that’s much harder if you’re dependent on outside vendors. For example, Tesla, knowingly or unknowingly, was a robotics company from last decade - the car is essentially a robot on wheels. As Tesla leaned deeper into owning their software, data, and pipelines, that’s exactly what enabled all these reinventions and new products that came out of it, most importantly Optimus, which is soon releasing. Engineers who built the system themselves understand it deeply enough to iterate fast and take real risks- that’s hard to get from a vendor relationship. Not against software vendors at all, it’s just a matter of choice some companies make when they want to keep owning and reinventing everything long-term.

[Amrit__ @Amrit__M]

@iam_smx Ig! Probably SpaceX is also building their own cameras fr.

[Amrit__ @Amrit__M]

You’re definitely right from your side, and fair to flag your own bias as a software builder — everyone’s raising concerns from where they sit. The real point is the problem exists everywhere, it’s just a question of scale and urgency. You believe better middleware should exist instead of everyone rebuilding it. On the other end, hardware and supply chain are the actual bottleneck for companies like Cosmic and etc. Both are true. But here’s the thing if the software solution isn’t fast enough to actually solve the problem when a company needs it, robotics companies will just build it in-house anyway, not because they want to, but because waiting isn’t an option. So it really comes down to speed: how quickly a real solution shows up determines whether “build vs buy” even stays a choice.

[Amrit__ @Amrit__M]

@sanskxr02 @thelcjones True!

[Amrit__ @Amrit__M]

Respectfully disagree with this framing. You’re right that status-seeking is endless, but that’s not what Elon’s actually pointing at — he’s talking about goods and services abundance, not human ego. Break it down fundamentally: money exists as a proxy for access to goods and services. If those become abundant enough food, energy, shelter, healthcare the thing money is actually a claim on starts losing its scarcity, and with it, its weight. Think about what that means for the millions still facing hunger and homelessness today: imagine every human eventually having what they need to survive, not as charity, but because production costs collapse toward zero. That’s a bigger deal than a handful of people chasing status forever. Money will still matter, sure, but by the time abundance actually arrives, it won’t carry anywhere near the weight it does today because survival and dignity won’t be gated behind it anymore.

[Amrit__ @Amrit__M]

Agreed, and honestly I think it’s that no one really wants to take on the hard things. Hardware is hard, it’s easy to give up on, and supply chain is arguably even harder than anything else — people should really notice that. Running Cosmic Robotics yourself, I can understand the pain and what you’re doing requires a lot of exactly what you’re talking about, and you’re definitely heading into a massively difficult problem worth solving. It’s just that right now, barely anyone’s actually trying to solve the hard problem. I pivoted into engineering myself over the last 6 months, both software and hardware, multidisciplinary. I may want to solve this problem too, honestly not totally sure yet, but I hope someone in the US starts doing this soon. I might start researching this myself starting today. Looking forward to what you and the team achieve — maybe I’ll even join as an intern at some point, hopefully. I love companies that do hard things, things that actually add value to the world.

[Amrit__ @Amrit__M]

True, no wind definitely helps. But finding and confirming a genuinely safe landing spot is still its own problem the Moon’s covered in fine dust and scattered rocks and craters, with no wind or rain to smooth things out over time like on Earth. Looking flat and actually being safe to land on are two different things.

[Amrit__ @Amrit__M]

@ishand @forwarddeploy @SpaceXAI Applied 3 days ago checking my inbox every hour!

[Amrit__ @Amrit__M]

Starship didn’t blow up today — it’s literally floating on the water like a boat. Genuinely surreal to watch. This wasn’t even expected to be this clean: @SpaceX called it their softest splashdown yet, ship swooped up from the bellyflop, landed upright, then just… tipped over and floated. This is probably the smoothest landing they’ve ever pulled off. Congrats to the whole @SpaceX team — this is truly insane, and I still can’t stop looking at the images of Starship just floating out there after a full flight.

[Amrit__ @Amrit__M]

I’d agree with a few things here — there’s definitely a lot of larping about revenue, and a lot of people are more locked in on raising a pre-seed than actually building something good for the world. It’s become so easy to raise a pre-seed these days, and founders end up larping more on LinkedIn just because they raised half a million or some other number, rather than telling the world what the actual benefits are. I feel people internally seek more money, power, and the feeling of having someone look up to them or being liked — that’s definitely a pretty good weakness to have. Anyways, there are still a lot of good founders and companies making good things possible. It’s really just about the purpose you have deep down, not the place.

[Amrit__ @Amrit__M]

Speeding is true, but that’s a settings issue — @Tesla could easily make FSD strictly follow speed limits by default or remove that flexibility altogether if regulators require it. On driver attention, the system already alerts the driver whenever eyes leave the road, and the eye-tracking keeps improving with every software update. And yes, roundabouts, lane changes, and intersections are complex maneuvers — but not for Tesla FSD specifically. It’s genuinely the safest self-driving software out there, and once it’s fully deployed, I believe we’ll see accident rates fall dramatically. And of course, you have the best engineers now and in the future as well — beating FSD is definitely not going to be easy.

[Amrit__ @Amrit__M]

I personally feel he’s already reached a level of winning the compute race and he’ll continue winning the compute race going toward ‘27. I also think his teams are very much trying to win the models race as well, but I’d say they’re targeting a whole bunch of different things that are valuable, rather than aiming for benchmarks like other models.

[Amrit__ @Amrit__M]

The trauma she’ll carry from this for the rest of her life is genuinely hard to even imagine. It’s hard to understand how someone capable of this exists in the same world as everyone else. Hoping she gets real support, safety, and full justice — and that whatever the legal system can do here, it does.

[Amrit__ @Amrit__M]

Fully agree — compounding doesn’t show up on a
Resume, it shows up in trust, reputation, and equity that only matter if you stay long enough to earn them. Job-hopping optimizes for how a career looks; staying and building optimizes for what it actually becomes. Nearly every extraordinary outcome traces back to depth over the appearance of momentum.