It's very nice to have written for so long... I often think I should write more for myself than for others.
HN user
po
Technical co-founder, currently living in Japan, previously living in NYC.
https://mastodon.social/@poswald
https://www.makeleaps.com/
I run Hacker News reader meetups in Tokyo.
Always looking for awesome people to meet up with. Get in touch!
pauloswald@gmail.com
[ my public key: https://keybase.io/poswald; my proof: https://keybase.io/poswald/sigs/HVlTjzK0n89ogL4IrN4QAX1JSJJUZRzfp_9MrlPIJ0A ]
as long as the business model is:
- users want the best/smartest LLM
- the best performance for inference is found by spending more and more tokens (deep thinking)
- pricing is based on cost per token
Then the inference providers/hyperscalers will take all of the margin available to app makers (and then give it to Nvidia apparently). It is a bad business to be in, and not viable for OpenAI at their valuation.
Did you read the comment you're replying to? It's talking about the DEI selection process being blind and instead focusing on outreach to get a more diverse input. You wouldn't be denied anything due to your sex under a system like that. It has nothing to do with what you're talking about.
Have you ever lived with or helped a person with AD? It's not cellular senescence. What you're talking about is fine and well, but AD is a devastating disease that has very particular symptoms. We may not know all of the causes, but reversing cellular senescence isn't going to solve this.
Researching and curing AD is not barking up the wrong tree. There is a horrible deadly monster in that tree that needs defeating. I hope people also get scientific funding for other age-related issues.
I happened to be walking a few blocks north and one east of this when it happened. I saw dozens of people running up the street in a panic. Office ladies running barefoot with their heels in hand. I asked what had happened and they said there was a giant explosion that could only have been terrorism.
The wikipedia article has a few more photos: https://en.wikipedia.org/wiki/2007_New_York_City_steam_explo...
The aftermath photo gives you a good sense of it: https://www.nbcnews.com/id/wbna20184563
Yeah I don't think CRM is the correct thing in this case... I just think that there needs to be some new set of incentives put in place such that the culture reinforces the outcomes you want.
Science needs an intervention similar to what the CRM process (https://en.wikipedia.org/wiki/Crew_resource_management) did to tamp down cowboy pilots flying their planes into the sides of mountains because they wouldn't listen to their copilots who were too timid to speak up.
...on the evening of Dec 28, 1978, they experienced a landing gear abnormality. The captain decided to enter a holding pattern so they could troubleshoot the problem. The captain focused on the landing gear problem for an hour, ignoring repeated hints from the first officer and the flight engineer about their dwindling fuel supply, and only realized the situation when the engines began flaming out. The aircraft crash-landed in a suburb of Portland, Oregon, over six miles (10 km) short of the runway
It has been applied to other fields:
Elements of CRM have been applied in US healthcare since the late 1990s, specifically in infection prevention. For example, the "central line bundle" of best practices recommends using a checklist when inserting a central venous catheter. The observer checking off the checklist is usually lower-ranking than the person inserting the catheter. The observer is encouraged to communicate when elements of the bundle are not executed; for example if a breach in sterility has occurred
Maybe not this system exactly, but a new way of doing science needs to be found.
Journals, scientists, funding sources, universities and research institutions are locked in a game that encourages data hiding, publish or perish incentives, and non-reproducible results.
Have many other projects put MPC servers (https://modelcontextprotocol.io/introduction) to use since it was announced? I haven't seen very many others using it yet.
Right... I think it's going to be way more than just data as well. You can expect all gov functions will start to show up in his app. Driver license applications, tax filing, etc...
My prediction is that any future data that is published will likely show up in the x.com app. I believe they are trying to privatize all government functions and Musk wants X to be the official government super-app similar to China's system.
This led to slower device performance without informing users, which is a removal of expected performance functionality.
As opposed to the device unexpectedly shutting down due to a degraded battery not being able to push enough energy to support the CPU? They didn't remove expected performance, they prevented crashes which are by definition 0 performance. All Li-ion batteries degrade over time. That's not removing a feature...
This whole thing was totally overblown.
I suspect this is related to why a string quartet is the right number of musical voices. Two violins, viola, and cello give you a very fulfilling number of separate ideas to track without overwhelming you.
Whether or not Zuckerberg himself is alive (he is), this version of him is long dead.
Yes, that was the joke that apparently wasn't obvious enough. :-)
Historically, those who controlled the media controlled the message. If you're the only one with a printing press, you control what people read. Same with radio. Same with TV.
But what happens when everyone can put their message in front of a lot of people? When the playing field is level? When everyone has a printing press, the ones with the best ideas are the ones people listen to. Influence can no longer be owned. It must be earned
Man, Zuckerberg would be rolling over in his grave if he saw what happened in the following years and our current engagement/algorithmically-driven media ecosystem.
I find this book to be a bit sad. I do believe they were trying to do all of this stuff but it definitely went off the rails.
yeah I looked at that too... and that aspect is useful.
However, I'm not really looking for a layer on top of ledger as much as I'm looking for a configurable web-scraping system (using the local password manager) that can be run to get the csv/pdf/etc.. files needed to create the ledger.
Plain text accounting is cool but I think one of the biggest barriers for people is downloading bank data into a standard format.
The banks are never going to embrace much more than CSV or excel files... the various data aggregation platforms (yodlee, plaid, etc...) are not open source or hobbyist friendly.
Back in ancient times there was a company called Wesabe (https://en.wikipedia.org/wiki/Wesabe) that wrote software that did bank syncing on your desktop. Mint.com basically put them out of business but I still think about that approach. I think it could work for open source.
Has anyone else?
Following the St. Petersburg attack, the Federal Security Service (FSB), in an event that may ring somewhat familiar to many in the United States and Europe, asked Telegram for encryption keys to decode the dead attacker’s messages. Telegram said it couldn’t give the keys over because it didn’t have them. In response, Russia’s internet and media regulator said the company wasn’t complying with legal requirements. The court-ordered ban on accessing Telegram from within Russia followed shortly thereafter. Telegram did, though, enact a privacy policy in August 2018 where it could hand over terror suspects’ user information (though not encryption keys to their messages) if given a court order.
...
... Pavel Durov, Telegram’s founder, called on Russian authorities on June 4 to lift the ban. He cited ongoing Telegram efforts to significantly improve the removal of extremist propaganda from the platform in ways that don’t violate privacy, such as setting a precedent of handing encryption keys to the FSB.
https://www.atlanticcouncil.org/blogs/new-atlanticist/whats-...
This one coins many new phrases that will stick in my mind. Phrases like "Big teddy bears" are actually useful in that they roll up a subtle or complex idea into a pattern that we can then talk about. Similar thing happened with things like "Gatcha mechanics" in mobile games for me.
The action isn't between political parties, it's the factions within the party. It's structured differently than the US so it looks strange that it's one party in power all the time, but there is plenty of push/pull for satisfying public opinion going on within the LDP. They have to _really_ screw up or ignore public opinion for a long time for the public to choose a different party.
Edit: come to think of it... can you imagine the leader of a political party choosing to step down due to popularity issues, allow a subordinate take the party in a new direction, and suddenly people are much happier with that party? seems oddly familiar. Imagine if the gov was structured in such a way that this is considered normal and then you basically understand why one party is always in power.
10 segment LEDs actually... the KO-II that this seems to be based on has similar ones but they're a little less curvy: https://teenage.engineering/products/ep-133
My 6 year old loves the colorful Sunday ones... but he has me read them to him and I don't think for many of them he doesn't 'get it' yet. My guess is that the truly ideal age is around 9 or 10. That being said, I think there's no wrong time to introduce them.
My sister just came to visit me in Tokyo and asked what to bring as a gift... I said, "If you can, please bring my Calvin and Hobbes collection for my kids to read." My 10 and 6 year olds have been devouring them ever since.
There is something electric and timeless about these strips. I am certain they don't understand all of the vocabulary but still they read it. It is a format that lures kids in and then uses that attention it has earned to stretch minds. Re-reading it as an adult also rings true in a totally different way. Calvin's parents become sympathetic compatriots.
It's smarter than most adults but captivates kids. It is a decade of work that deserves all the awards that could possibly be given.
There are a lot of doubters in the threads here and I think a lot of it has to do with the generative art that accompanies the article. They haven't picked a technology and it may actually end up being driverless cars on dedicated lanes, or automated rail technology. The point is non-stop driverless tech. The country is very well aware of the coming labor shortage and is going to rely heavily on automation in the future.
Anyone who has ever stepped on a Shinkansen can understand that Japan has a way of getting shit done. It may start a bit later than other countries, but when they finally decide to execute (and this is still in planning stages so it might not), it happens surprisingly quickly.
I've watched them totally revamp Shibuya station over the past 10 years in ways that make it utterly unrecognisable to someone who hasn't been here since then, all while never taking down-time. There is another article on the HN front page titled "Why it takes NYC nearly 10 years to install 500 feet of pipes" right now... that's perhaps coloring people's expectations of the possible.
I agree with you and I'm a bit surprised that more people don't see the difference between what Manning and Snowden did and what Assange was up to (including apparently Snowden himself).
At the time, I was initially a person who thought that what Wikileaks was doing was a net good for the rule of law, but changed my mind when I learned about the selective nature of what they publish. The fact that they were playing politics, pushing conspiracy theories, and actively coordinating with the Trump campaign completely discredits any moral high-ground they had. You can say that what happened to him is unfair and that may even be true but Assange is no hero.
This is great advice and I've actually built ponds on this principle in the past. It works well. There are tons of Youtube videos on this subject as well. The channel ozponds is pretty good for this.
What I am now interested in is the construction of spring boxes and finding and digging a natural spring from a wooded area. I used to do this when I was younger, but I've found that while the internet generally has all kinds of new techniques that I wasn't aware of, unfortunately, I haven't found a great source of info for this topic yet.
Thank you for looking it up!
Is there a name for the practice of embedding a completely unrelated video into the middle of an article? I find this practice to be so mystifying. Does that work on readers?
How would we get there from here? I’m not smart enough to figure out what the regulatory regime is that would ban most of what private-equity does and tilt the playing field in favor of resilient lifestyle businesses.
One good way is to pay attention to the details, learn how things are made (YouTube helps with this!) and not be impressed by shitty work, in general. I buy plenty of MDF-built products but I also know the difference. There are just so many people who have no idea.
If JPY appreciates against USD they don't make money or can even lose money. I'm sure there is some hedging going on, but I don't know for sure the mechanics of it all: https://www.thebalancemoney.com/yen-carry-trade-explained-pr...
This is exactly right, and also demonstrates why the Japan is in a bind.
They don't mind the currency going low as it helps exports but going too low, too fast is a problem. Increasing interest rates to narrow the spread will affect the economy. So they are walking a very tight line here between JPY/USD spreads and domestic economy/interest rates. All global economies are to some extent but since JPY is trusted and has had low interest rates for decades, it's popular within the finance industry for this kind of trade.
As a retail investor, you can do Forex trading or you could do time deposits at a Japanese bank (assuming you live here). SMBC Prestia is offering 6.5% on it right now for example:
https://www.smbctb.co.jp/en/timedeposit2404/?icid=24_en_top_...