David Drake

Co-Founder @tectaai
I like to talk about Sports, Science, & Stocks

[David Drake @David6Drake]

@jukan05 $QCOM 🙏

[David Drake @David6Drake]

@iamrahstradamus I think Tank is a bit closer in weight to those guys, but yeah true that’s a good point. I just think Canelo was at a point in his career where the risk/reward was not there to take on a challenge like that. Tank also got clowned for instating a rehydration clause with Ryan

[David Drake @David6Drake]

Interesting

[Andrej Karpathy @karpathy]

One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anything goes, full stream of consciousness. Sometimes I declare it up top, something like "switching to speech recognition sorry for any typos...". Sometimes I turn it into a small interview of a few turns. But I find that the LLMs are somehow very good at reconstructing long incoherent rambles and often their echo of your own tangle of thoughts comes out quite a bit cleaner than what you started with. The result is that you improve the mind meld and have to correct things less from that point on.

[David Drake @David6Drake]

@iamrahstradamus Benavidez is huge. Why should Canelo have to fight a guy who is realistically two weight classes above and in his prime, when Canelo is a bit past his day. Canelo learned his lesson when fighting Bivol. Canelo fought legends, you can’t hold not making that fight against him imo

[David Drake @David6Drake]

This is why I hate stocks

[TrendSpider @TrendSpider]

GOOGLE JUST ABSOLUTELY SMASHED EPS 🔥

Q2 EPS: $9.11 vs $2.90 est
Q2 REV: $119.796B vs $116.817B est

Operating Margin 34%
Google Search Revenue $63.27B
Net Income $112.107B
Advertising Revenue $81.629B
$GOOG $GOOGL

🟥 -0.85%

[David Drake @David6Drake]

Great explanation of this misconception!

[Patrick C Toulme @PatrickToulme]

I am seeing a lot of misinformation online on open-source model cybersecurity and the risk of data loss when using open-source models. I want to briefly explain why these claims are factually wrong and extremely misleading.

First some background information:

What is a model?
A model is fundamentally a set of weight matrices. These are floating point numbers. Billions or trillions of these raw floating point numbers.

What is an open-source model?
An open-source model means the model provider (example Kimi, GLM, Qwen) publishes the weight files on a site like Hugging Face. This is a raw bin file of floating point numbers. Modern releases use the safetensors format, which does not execute code.

So why is the claim “open-source models steal your data” wrong?

When you load an open-source model on your own GPUs or TPUs in your datacenter - zero data ever leaves your environment. This is the entire point of open-source weights. You never call a model provider API, and data never leaves your environment. You do not even need an internet connection to do this. You have the model locally on your disk, and you have the chips in your own datacenter.

The same is also true for when you use open-source models on cloud providers like Azure, GCP and AWS. The model is hosted in cloud datacenters. No model provider can see or access your data, and data never leaves your cloud tenant.

When you see people claiming open-source AI models are a cybersecurity risk for data loss please ask for a community note, as this is pure misinformation. Feel free to ask questions below.

[David Drake @David6Drake]

@Venu_7_ When Gavin speaks, I listen

[David Drake @David6Drake]

Not only is Messi the GOAT of football, but might be the greatest athlete of all time. That ball couldn’t have been more perfect!

[David Drake @David6Drake]

@StockSavvyShay Couod be huge!

[David Drake @David6Drake]

@AravSrinivas What about nuclear?

[David Drake @David6Drake]

@SchillWaren @DailyFaceoff That’s the stupidest thing I’ve ever heard. So is the best all time scorer from Kazakhstan also better?

[David Drake @David6Drake]

@unusual_whales They pretty obviously attempted to only release enough to satisfy the public, but not expose anything too important

[David Drake @David6Drake]

@elonmusk This is wild to think about

[David Drake @David6Drake]

@PolymarketMoney Wow

[David Drake @David6Drake]

@MMatters22596 If this happens, I’ll be a rich man

[David Drake @David6Drake]

@elonmusk Excited to try it out!

[David Drake @David6Drake]

@danielnewmanUV Facts

[David Drake @David6Drake]

@StockSavvyShay $ROK

[David Drake @David6Drake]

I will love Kawhi forever, but giving up those picks for a guy who breaks down every year seems like a mistake to me

[Shams Charania @ShamsCharania]

BREAKING: The Los Angeles Clippers are nearing deal sending Kawhi Leonard to the Toronto Raptors for Brandon Ingram, Gradey Dick, 2 first-round picks, 1 pick swap and 2 second-rounders, sources tell ESPN. A return to Canada for the Raptors champion and two-time Finals MVP.

[David Drake @David6Drake]

@ThePensInsider Would make sense if there was a pick coming back as an incentive to free up Vegas’ cap space, suprised there wasn’t one