Why is xAI giving up their advantage? Is this a signal that their frontier model improvements are plateauing and decided there is no value in hoarding all their compute?
xAI have (with some questionable ethics) managed to actually build a data centre, so they have a ton of compute but not much inference demand for their model which is second-tier.
Everyone else trying to build data centres is really struggling (turns out building physical things is not as easy as writing code, who knew).
So Anthropic have the model but they're compute starved because everyone else they've signed agreements with still mostly have piles of dirt and the GPUs are still in Nvidia's warehouse somewhere.
It's a bit of a win-win: xAI's financial numbers will be massively improved by the revenue using otherwise useless data centre capacity, Anthropic get the compute they desperately need albeit probably paying through the nose for it
Elon got rid of most of the core xAI researchers recently. It looked like firings and he said something like Grok needs to be rewritten because it was fundamentally built in a bad way. My guess - Grok was not one of the top two models and had no traction beyond users on Twitter - so investing in it did not make sense. Or maybe he needs time to build a new team and rebuild Grok, and while he’s doing that, he needs revenue to make up for all the GPUs they bought.
Either way, I doubt he has any ‘real’ advantage in this game. It’s OpenAI and Anthropic, and then everyone else including the open weight models.
Boy I’d love to have the sort of ‘no real advantage’ that got me $1.9bn in revenue per month from Anthropic.
Make no mistake - Elon’s charging premium pricing because he has something Anthropic realllly needs. In exchange, he will get significant revenues added to the balance sheet along with the cursor team right around IPO. It’s very good business for Elon and SpaceX.
What advantage? Has there ever been any indication they’re leading in any segment? Sure Elon has thrown a bunch of money at hardware, but to what end?
And frankly as bad as Altman is from a: if AI is really going to disrupt humanity do I want this guy in charge? Elon is 10x worse. So why would the best and the brightest ever work for him?
yes, but it's only advantage if one is compute-strained and the other isn't. if they both have lots then there's no advantage. if one doesn't fully utilise their compute then it's not an advantage either
Well, other than your ability to turn that into cash by renting it out for the highest price to someone who needs it, you can promise prospective employees that are supposed to use that infra to train models that they won’t be compute starved.
You can kick off more model training runs and experiments than your competitors.
You can kick off a $1-2t IPO claiming you are going to capture a large portion of the largest TAM the world has ever seen.
They have neither the most resources nor the best models. They are mediocre at everything except the CSAM generation market, they've got that one cornered
Could have (past tense). They now have rented the capacity out, presumably with a multi-year contract.
A rental deal, even at a premium price, actually makes great sense for Anthropic. Rather than building themselves, they are de-risked by renting instead, and they don't own the depreciating asset. They've also compute-starved X.ai as a potential competitor as a marginal added bonus.
I see this as a huge warning sign. If a frontier ai lab is in this position of renting their own capacity, imagine how much overcapacity there is in the system.
Outside of VC money, and circular financing, the only external money coming into ai are into open-ai, and anthropic via their subscriptions and APIs.
I think any warning this provides is more about the ability to forecast demand. Obviously it was a giant leap of faith, that didn't pan out, for Must to build these data centers without having a product with a growth trajectory to more rationally justify it.
OpenAI and Anthropic (and their investors!) are obviously also out on a limb here, both with deals in place to 10x their available compute (roughly 1GW -> 10GW) over next few years. Maybe these growth estimates are a bit more grounded, but they have all sorts of assumptions baked in. I still would not be surprised to see OpenAI self-implode, and to see Google and Microsoft as the eventual winners here.
For sure. There’s also this kind of interesting meta strategy question: if you think you can float the huge capex with some internally valuable processes as a base case then you might just go really big. Dario and Elon’s temperaments come in to play here
They were briefly SOTA on some benchmarks, although there were suggestions there might have been some massaging of the results since real world usage showed lackluster performance compared to the benchmarks.
Musk's goal with X.ai from the start seems to have been primarily to compete with OpenAI, and given that Grok itself isn't doing that (not even mentioned in competitive benchmark tables), and his data centers are therefore mostly sitting idle, this deal with Anthropic helps him achieve that goal by helping OpenAI's #1 competitor.
Note how the initial deal was only for the older Colossus #1 data center, and now after losing his lawsuit against OpenAI the Anthropic deal has been expanded to Colossus #2 also. Coincidence?
Here is my template from obsidian that I use for my weekly reflection - customize reflection based on your values and priorities. I have goals to improve work-life balance, social connections (social isolation was a factor driving poor outcomes, and through deliberate consistent effort I have solved this problem).
Hyper scheduling: https://dev.to/maxpatiiuk/series/32301 (I stumbled upon this and implemented a form of it, although mostly I just like the colors in my calendar)
- [ ] Review year compass
- [ ] Review journal entries from the week
- [ ] Review last week's reflection
## Quick summary
> *Headline for the week*:
## Basic planning
- [ ] Set up outline of the week in Outlook
- [ ] Plan a fun weekend activity:
- [ ] Plan to visit one new restaurant:
- [ ] Plan one meet up or social activity:
- meditation (I combine with a fancy LED face mask to help reinforce the habit via my desire to combat wrinkles and acne - the cryoglow is better at acne than wrinkles so far)
- exercise (you can add notes)
- evening leisure time (if I don’t have dedicated leisure time, I end up revenge bedtime procrastinating/doomscrolling)
- stretching (there are two simple band stretching exercises that solved what I thought would be life-long neck and shoulder pain)
When I really struggle with productivity, I find the pomodoro system is a good bootstrap, and TickTick makes it easy to start. I like seeing the pomos on the built-in calendar.
I have a ton of respect for your approach. That said, as someone without ADHD, it seems somewhat odd that an inability to kick off executive function would be well addressed by adding an additional activity that requires executive function. Like, if I had to plan my day out with this document before doing things, I think I'd grow to dread the process, and be even more stymied - i.e. if it was hard to go clean the kitchen, why wouldn't it be hard to go write my dayplan?
Yet, I do hear this sort of thing works for people. I'd love to know more about what you experience and why this helps.
TickTick gives me reminders to do some of these things - daily reflection is a “habit” and weekly reflection is a recurring task. That helps me not forget. There’s still value in doing the daily plan regardless of whether I do it right after I wake up (things are going well) or if I do it many hours later when I realize my day has not gone well and I want to get back on track.
The calendars and checklists really help with not forgetting things, and getting back on task once distracted. I can have 100 adhd moments; the system can’t prevent that, but it can help me find my way back to shore when I’m lost at sea.
I don’t really struggle with kitchen cleaning, but sometimes I do set out to clean the kitchen and end up folding laundry or scrolling instead. I can do that but if I haven’t checked off “clean the kitchen”, I know to come back to it.
The weekly reflection is a chore but I set aside time for it, and I keep doing it because it works. I can spend an hour doing deep reflection, or I can rush through it, there is value either way. It is really just a check list, and check lists are very ADHD friendly.
Separately, I find that physical and mental health improve performance regardless - so a system that improves these factors _is_ a system that helps with ADHD. Getting to bed on time, and building the system that produces that outcome, is an ADHD-friendly system. I struggle with this, but I try to get better over time by experimenting and adjusting.
The reality is that the system will not solve ADHD problems, the system is just a tool - you still have to do the work. Same for the app that is being shared.
I have ADHD & use a very similar kind of daily notes template. I think the reason that it works for me at least is that it shortcuts some of the "not knowing where to start" executive function issues at the start of the day. I only need to remember one thing: open a new daily notes template. The checklist itself chains other good habits off the back of that, reminding me to e.g. make sure I've taken my meds, make coffee, put music on, think about what I've done recently & if there's anything to pick up again or dive back into & what I'm going to do for the day etc.
Yes, it can feel like a chore at times (especially forcing myself to do more in-depth periodic reflections) but without a system like this, I quickly find myself completely rudderless & my mental health & productivity both collapse pretty quickly.
You'll actually hear quite frequently that people diagnosed relatively late in life went under the radar for so long exactly because we developed our own massively over compensating systems like this & can appear (externally at least) to almost have the exact opposite of "executive function issues"!
Looks like joy-cons will have 'mouse-like' functionality and there's a 'C' on right joy-con but its functionality is not reveled. New Mario Kart showcased would probably be one of the first exclusives.
A few details are quite different from 8, notably the boost and character animations, it's definitely a new game.
Marketing will be difficult, MK8 already peaked graphically and has 96 tracks, and will still work on Switch 2. I hope they'll find real selling points for MK9.
Would have not surprised me if it's actually Mario Kart 8 2. (Technically that's already what Mario Kart 8 Deluxe is, so, actually, it would be Mario Kart 8 3).
I mean, at this point it makes little sense for them to start from scratch, releasing a newer game but with much less than the enormous amount of content provided by MK8D + DLC would seem like a very noticeable downgrade, so just revamping the old one would be a practical move, though I don't think fans would be happy with that.
MK8 was mostly flawless gameplay wise, how can it be improved? But at this point one has no choice but to trust Nintendo's ability to come up with surprises.
There are certainly some ways they can, I'd love to see a 100 man race or something crazy like that.
MK8 was also an iteration on MK7, with refinements to the handling, the addition of anti-gravity, and tweaks to items. It's certain there's going to be _some_ sort of mechanical shakeup.
Mario Kart sells like hotcakes; I doubt they'll have to do much to convince people to buy a new one, particularly folks who've played the old one for hundreds of hours.
yeah, i agree on that, makes more sense to update 8.
BUT, i don't know if i would use that as the first look at the new console, basically looks like really similar to a game that was released 10 years ago, i wouldn't buy a new system to play again mario kart 8.
I thought they were showing the retro compatibility feature, since the gameplay comes after the message that switch 1 games would be playable on 2 (maybe upscaled or something)
Everyone has a new design, maybe I'm more familiar with my Marios than most but I could tell immediately it has a more cartoonish design, and characters have a rubbery kind of stretch and bounce to their animations. You can see it notably on the closeup of Mario where he hops into a drift.
The art style is somewhere between the 2010s bog-standard Mario and Super Mario Bros Wonder.
To be fair, if we’re going by track alone, there’s nothing to say it’s not just a new track for the Switch 2 release (or even just released at the same time, but available on both).
That's so cliche and cringe nowadays, but the reason they didn't wait to do that is probably because of all the leaks. The specs, the name, photos of the console and internal components all leaked. Even the fan renders people were making turned out to be pretty damn accurate (https://www.reddit.com/r/NintendoSwitch/comments/1i008os/nin...)
Calling anything “cringe” is pretty self-referential. This slang just makes me imagine a bunch of genZ folk wincing nonstop with the heads in their phones. Must be exhausting.
As long as the internet has existed, we have been lampooning corporate keynotes. The gaming industry does this every cycle, trying to hype up incremental updates as if it’s the best thing to ever get released. See you again in a few years!
Used lunchmoney for years before moving to CoPilot couple of years ago. Can vouch for both of them to being great. I just moved to CoPilot because they got macos and ios apps unlike lunchmoney (web only back when I used it)
reply