I have been working on a language learning app for myself, and I am using a textbook that I like as the basis for an Anki inspired “learning tree”. This is working pretty well because I can build progressions from the original table of contents.
Yes, generally if you get a decent progression from an LLM, I find it's copying directly from a source, like a book or course. Giving it a progression helps a lot.
A huge chunk of SoaceX value in their filing is attributed to their AI technology (aka Grok). I believe it’s 90% or more… Now, it seems they’re leasing the infrastructure required for Grok to scale to Anthropic and Google. I wonder how that math works…
But what is xAI? I thought that was the company that had the compute + Grok, the AI company? Since when does SpaceX (which I thought was a space company?) own AI-compute hardware and/or can do model hosting? Are all of Musks companies just one big thing now where the names no longer matter, or how is it supposed to work?
Edit: seems I'm just a bit behind: "xAI — now part of SpaceX ", seems really strange for a space company to buy an AI company, but I guess rather that, than the other way around.
What people are describing here is the colloquial definition of "server", being compute hosted in space. Not a broadcast signal that is interpreted on Earth.
Musk sold Twitter into xAI which he then sold into SpaceX as a financial engineering effort to lessen the impact of massive debts and cash burn. The IPO and some clever structuring is the final step in the process.
Not really strange... if the goal is to go to mars, you probably need robots, those need intelligence -> ai. It fits pretty well, especially because you want to own all the core technologies as a company.
Why 4-5 companies instead of one then? I thought the goal of SpaceX was to get to Mars, why does xAI need to have that same goal? Or he didn't think xAI was suitable for that goal, then changed his mind so merged the companies?
You are overthinking this. The whole purpose of the SpaceX / xAI merger is for Musk to launder his failing companies to make them more palatable to the public. Not unlike the complex Mortgage Backed Securities of the GFC era which had a ton of low quality debt but yet were somehow assigned spotless credit ratings. Twitter is also being rolled up into SpaceX for the same reason.
The stated goal is to "go to mars", the real goal is to make money.
He sold his failing but hype business to his soon-to-IPO successful but kinda boring business.
It's a way of laundering the debt and dumping into investors as he pitted different indexes against each other to force his way into one of them, and have people's 401k buy into them. Its a ton of money.
I wouldn't be surprised if Tesla is bought into spaceX in the future.
It has nothing to do with Grok, at least not the current iteration. SpaceX is the only company that can concievably launch large scale orbital compute.
I’m out of the loop, why is compute better /after/ being launched into space? Is the idea just to be co-located within the ISP to reduce round trip time to the LLMs?
There would be some benefits, assuming you could do it for a reasonable cost. For one, you have effectively uninterruptible power using solar panels in space. And it's free, too, once you have the hardware in place.
And you don't have to deal with any of the site selection stuff you have for terrestrial data centers. No NIMBYs. No politicians trying to extort bribes. No water problems.
In space there are no earthquakes, tornadoes, or floods.
I'm still skeptical. It's hard to believe it costs so much to build a data center on the ground that putting it into orbit is an economically viable alternative.
Yeah, launch costs alone make it infeasible, and power being "free" exacerbates the cost (gotta get all those panels up there). Cooling is also dramatically harder, plus shielding, and it makes repair/upgrade basically impossible.
I'm not going to assert that large scale space compute will never happen, but I feel confident saying it won't happen this decade or next.
When they get the kinks out of Starship, launch costs will be dramatically lower than we're used to thinking. They'll probably be using it to launch Starlink sats in earnest next year, so I don't think that would be the long leg.
I used to think heat would be a problem, too, but I've come around. It's a consideration, but it's doable. Remember we already have some pretty high power sats up there, so it's not something we haven't already been working around.
IMO the big cash drain will end up being maintenance, as in, you can't do it. If you have a box or a power supply fail on the ground you can swap it out. Anything in orbit would have to be replaced.
Things to think about: the russian satellites currently already in orbit and being tested to jam global GPS signals: https://youtu.be/tz23G_UXCGA?si=Jkg7hYnwER39-FXf (Veritasium). In all these sci-fi space datacentre scenarios, has anybody considered that astrolaw is in no way currently equipped to deal with datacenters in space? Neither are international relations.
Last time I did the math, launch costs were well balanced against permitting delays (mediated by interest rates). The break even rests almost entirely on radiator mass efficiency (which is, admittedly, a function of launch costs).
Like, if everyone’s terrestrial datacenter projects start getting blocked, and demand for AI continues, the price a rational buyer would pay for in-orbit compute could get ridiculous enough to break even on current kit. And current kit in launch vehicles, radiators and solar panels is advancing.
I don’t think the thesis is met yet. But it’s less ridiculous than I thought it was before I sat down with pen and paper.
Wouldn’t it be easier and cheaper simply to focus building your data centers in other poorer countries where your permits won’t be blocked? Why would the next most logical choice be “let’s put it in orbit”?
Neurotone.ai || Remote (Americas East / Western Europe) || Full-time || neurotone.com
We build AI-assisted auditory training for people with hearing loss. Our product is live, in clinics, and growing internationally.
Backend Engineer: Postgres-heavy work on Supabase: PL/pgSQL functions, pgmq queues, Deno edge functions, billing and cron systems. You like databases enough to enjoy writing triggers and RLS policies, and you want to own systems end-to-end rather than just ship tickets.
QA Specialist: Manual testing of our iOS/Android app and web tools, working alongside our head of QA on a clinical product where data quality, platform stability, and accessibility are critical to our mission.
Note: Candidates must be in time zones no further west than EST; Western Europe is fine.
If interested, contact travis AT neurotone.com. Join us in making a difference!
Hi! I'm Her and I have 6+years experience as a QA. I'm currently work as an Agile Delivery Lead and QA Lead for my clients based in AU, EU and UK. I've also worked with tech companies and startups spanning from different industries (Fintech, Banking, Mobile Apps, AI, Cloud) including IBM. I would like to bring my expertise to help you in your company provide quality products to your everyday users. Looking forward to working together!
I think we've developed software with "ROI" in mind for so long, that by now most people forgot how it was to use devices and interfaces that were made with passion and by taking your time, experimenting and finding the right way, rather than just rushing through stuff and optimizing everything for money.
I remember Flipper Zero had a ton of doubters early on too, myself included. I think I'm now willing to give them more slack to actually experiment and create something even more ambitious, as they successfully executed it the first time most doubted them.
I've worked in startups long enough to see many founders build without considering ROI.
It's not rare at all.
The reason you don't see those projects is because they don't make it very far. Big projects take a lot of effort and people and most people expect compensation for their effort. You can't compensate them without ROI.
As an open-source project they have some benefit of getting contributors to do some of the work. The hardware still needs ROI to exist. Making those custom parts requires up-front capital, which is going to need ROI to pay back.
Here’s what’s crazy: by making this widely available to SMB, they will soon have enough training data to beat this benchmark —- in probably less than a year is my guess.
I tend to agree, but the gains will come at the expense of the early adopters. Then again, this has been the case in so many industries throughout history.
Overkill in what way exactly? The LOC of the project shouldn't have any bearing on most people's usage of the project. SQLite is one of the well tested and mature projects in the world. What exactly would motivate someone to use PeakSlab instead? What problem are you solving?
I'm solving a simpler problem. Just making cross platform dictionary progressive web apps with indexes and full text search and HTML tags and uppercase letters inserted back into the text on render so they don't interfere with search.
SQLite is 1.2mb in combined wasm and JavaScript and not really designed for my use case, so I would have to add all the things i need anyway like compression and HTML tag insertion. For my use case which is just for pwas SQLite takes too long to load and the files are too big and the search isn't tailored. So I made something else in 38kb instead
We can thank RFK for anti-vaccine leadership leading to changes to the childhood vaccine schedule which make preventable diseases harder to stop. You can celebrate more cures all you want, but as most people know, prevention is the best medicine, and the guy is lukewarm on one of our best defenses when it comes to keeping kids safe (vaccines).
And, let's not forget that RFK said: “every Black kid is now, just as a standard, put on adderall, [selective serotonin reuptake inhibitors], benzos, which are known to induce violence. And those kids are going to have a chance to go somewhere and get re-parented, to live in a community where there’ll be no cell phones, no screens, you’ll actually have to talk to people."
So yeah, according to RFK, every black kid in America deserves "re-parenting". He should resign today.
reply