bjorney

The poster above asked for a use case. I gave one.

Frankly I don’t give a shit if the market penetration of said use case doesn’t meet whatever arbitrary cutoff you have deemed sufficient for something to “exist” or not - the QR code on the back of every north american bag of Starbucks beans is proof enough. Whether its more or efficient than a traditional RDBMS is irrelevant

Once again, we are talking about blockchain, not Bitcoin

You realize blockchain is used by many large companies for practical purposes, not just by hobbyists swapping magical internet money, right?

Many large retailers (e.g. Walmart) and pharmaceutical companies use managed blockchain solutions (e.g. IBMs supply chain software) to track end to end process flow and see the pedigree of products at their end destination, because it means the end user doesn’t need to request unfettered access to 6 different companies ERP systems to know when the hell their purchase order is getting delivered

The valuation of Bitcoin is a completely separate topic than practical use cases of blockchain.

It’s a solution that allows two parties, who are so paranoid they don’t trust banks, let alone one another, to send funds and maintain a record of transactions with one another.

To be fair it’s power consumption is effectively zero at standby and only 4-5W at idle/light usage.

If you were worried about this amount of power usage you would be better off unplugging your microwave when not in use to avoid running the clock display

also for the environment, I would think. It saves a ton of useless traffic

GPT is worse and it’s not even close.

My PC can serve up a hundred requests per second running an HTTP server with a connected database with 200W power usage

It takes that same computer 30-60s to return a response from a 13B parameter model (WAY less power usage than GPT), while using 400W of power thanks to the GPU

Napkin math, the AI response uses about 10,000x more electricity

AFAIK if you spend at least 2 years studying here you automatically qualify for a 3 year work permit. I think rolling that into permanent residency is a lot easier than just applying for a work visa or PR out of the gate

International student tuition is way more expensive here in Canada than it is for citizens, but I’m not sure how it stacks up against normal US tuition.

Grain of salt, everything I’ve said is based on anecdotes from people I know who went through it

Glass will absorb and retain more heat for longer;steel will absorb energy and heat up more quickly, and dump it just as fast.

Which was my point - 400g of room temperature ceramic is going to absorb way more heat from 250ml of boiling water than would be lost from the glass-air (or even steel-air) interface during the 2 minutes it takes to do a pourover.

If both cones are preheated thoroughly, yes, the steel cone will shed heat faster, however I feel like this is also negligible compared to evaporative heat loss and subsequent transfer to a cup

stainless has ~10x the thermal conductivity of borosilicate glass

Glass has double the heat capacity, and I would assume greater mass due to thicker construction. So unless you are preheating fully to boiling temps first every time, there will be more heat loss to the glass over the course of ~1-3 minutes

Or if you are on a Boeing plane and a side panel/door spontaneously flies off off you don’t get sucked out

/s, but not really /s

“making me look bad doesn’t make you look any better”

Do you think an open Hario switch basically IS a v60?

It is. It’s just a glass v60 with a seal at the bottom

Abolishing the monarchy would involve rewriting the constitution - if that was happening every province would want to slip in their own terms - Quebec would want specific French language rights and autonomy and if Quebec got their way Alberta would want something similar. We successfully altered the constitution back in 1982 - it took 2 years and the country almost blew up over it.

Basically it would be a total shit show. Considering the impact the monarchy has on our day to day life (basically zero) it’s easier to just let sleeping dogs lie

Murica.

This was literally the overarching plot for the last season of curb

https://youtu.be/dHIPXbLsY_Q?si=KG-IWg7GTeqQ8jiT

It’s really not that complicated. At a high level:

• $5/mo for having the service turned on
• $5/mo for every TB storage above and beyond the first 1TB
• $1 for every TB of data transfer beyond the first 1TB in a month

And then divide those numbers because it’s actually billed by the hour

deleted by creator

The word ‘decipher’ is doing a lot of heavy lifting. I’m wondering if they socially engineered or just found it written somewhere in the house?

You can plausibly brute force up to 4, maybe 5 words of a seed phrase. It takes longer than a normal password because every seed phrase is technically valid, so the only way to know if your brute force is successful is to generate thousands of addresses at each of the different derivation paths you may expect funds to exist at.

The same seed phrase is used for Bitcoin, Ethereum, Monero, etc, but each currency uses the seed phrase to generate addresses in a slightly different standard. Additionally, each wallet uses a slightly different variation of that. Within each wallet is a notion of accounts, and within each account you could have dozens of addresses. You need to generate each of those addresses, and scan each cryptocurrencies blockchain to see if those addresses have ever been used.

Realistically one of three things happened: his seed phrase was written down and they found it, it was password protected or on a drive with weak AES encryption and they cracked THAT instead, or finally, he used a hardware wallet and they exploited a firmware vulnerability to lift the PIN and transfer out funds and/or read the seed from the device

Does Google Cloud not count as “own hardware” for google?

That’s why the bars are so different. The “cloud” price is MSRP

That’s not really machine learning though. If you wanted to go way back, AI research goes back to implementations of hebbian learning in computer science back in the 1950s as a way of emulating human neurons. I was merely pointing out that AI was a computer science “dead end” until restricted Boltzmann machines were revisited by Hinton et al back in 2008 or so, and that 99% of the growth in the field has happened since the early 2010s when we reached a turning point where deep learning models could actually outperform classical statistical models like regression and random forests

in the coming decades

Given that in the past 15 years we went from “solving regression problems a little bit better than linear models some of the time” to what we have now, it’s not unfounded to think 15 years from now people could be giving LLMs access to code execution environments