Imagine thinking Warren Buffett could use “inside information” to make an investment he then holds for 40 years.
Props to the Buff-dawg for getting away with it for so long.
Don’t wanna sound bitter but I’d genuinely be disgusted to win a league title like this. How do their fans smile at this? It’s surely up there with the most uninspiring ways to win a major title. I’ve never seen a team more happy with the ball in the air than on the ground.
Probably one of the most severe flushes I’ve ever seen on alts, I didn’t even imagine alts had this much leverage in them. It feels like someone got hit very hard and will see a large body float to the surface soon, reminds me a little of summer 2021.
Good reminder to myself to own things that I am actually bullish on, and not things I am trying to shift on momentum. Some charts look like they’ll never recover, whereas some things look buyable for the first time in a while.
When everyone is making hilarious amounts of money I am always tempted to start using leverage again. It is almost impossible to fight the feeling that you’re not making enough, or everyone else is outpacing you. Good reminder that fighting that feeling and avoid the wipeouts is worth it in the end.
Check on your friends, likely a bad day for many.
Personally, am concentrating my bags into the things I am happy to own for the next few years, and shedding the fat. Realised I own some assets based on not wanting to miss out, rather than on some actual thesis. Days like today are much easier for me if I think my bags will bounce back, and much worse if I’m losing money owning things I don’t even believe in.
Don’t let a leverage blowup dictate your long-term views. The future is bright, good things to come, patience is rewarded.
嵐の後
Launch is getting closer.
Everyone who retweets this will get Early Access.
The number ONE collectables store for Pokemon with crypto payments is getting close.
The Pokemon Index outperformance will simply continue.
👀🔥
Just joined the @GoBundled waitlist! 🙌 It’s like Costco for subscriptions—big savings, one-click cancellations, extended trials, and more privacy! Check it out at https://t.co/99CcDmmR4a and follow them to get in early. #Bundled#SubscriptionSavings#GetInEarly
DeepSeek (Chinese AI co) making it look easy today with an open weights release of a frontier-grade LLM trained on a joke of a budget (2048 GPUs for 2 months, $6M).
For reference, this level of capability is supposed to require clusters of closer to 16K GPUs, the ones being brought up today are more around 100K GPUs. E.g. Llama 3 405B used 30.8M GPU-hours, while DeepSeek-V3 looks to be a stronger model at only 2.8M GPU-hours (~11X less compute). If the model also passes vibe checks (e.g. LLM arena rankings are ongoing, my few quick tests went well so far) it will be a highly impressive display of research and engineering under resource constraints.
Does this mean you don't need large GPU clusters for frontier LLMs? No but you have to ensure that you're not wasteful with what you have, and this looks like a nice demonstration that there's still a lot to get through with both data and algorithms.
Very nice & detailed tech report too, reading through.