flashman

That's funny because if I was running SIGINT at the NSA I would do all of the above, and also compromise Tor

"Making AI that is smarter than almost all humans at almost all things will require millions of chips, tens of billions of dollars (at least), and is most likely to happen in 2026-2027."

What an incredible claim to just slip in two-thirds of the way through! Where have I heard it before?

1965, H. A. Simon: "Machines will be capable, within twenty years, of doing any work a man can do."

I imagine that argument is a pretty easy sell. Of course governments will want to create moats that protect their incumbents.

AI models and their outputs aren't copyrightable because they're not made with human creativity :)

That's a coincidence, Australian astronomers published a paper on a new interferometry-based system for fast radio burst detection today too:

https://www.csiro.au/en/news/All/News/2025/January/Australia...

https://www.cambridge.org/core/journals/publications-of-the-...

> Next up: 1,156 prompts censored by ChatGPT

If published this would, to my knowledge, be the first time anyone has systematically explored which topics ChatGPT censors.

Jevons was talking about coal as an input to commercial processes, for which there were other alternatives that competed on price (e.g. manual/animal labour). Whatever the process, it generated a return, it had utility, and it had scale.

I argue it doesn't apply to generative AI because its outputs are mostly no good, have no utility, or are good but only in limited commercial contexts.

In the first case, a machine that produces garbage faster and cheaper doesn't mean demand for the garbage will increase. And in the second case, there aren't enough buyers for high-quality computer-generated pictures of toilets to meaningfully boost demand for Nvidia's products.

Yeah, GME for instance shows you can only think about it correctly in retrospect (or perhaps with perfect knowledge of all players' intentions).

what people are neglecting about Jevons is that fuel efficiency led to increased use of coal because you could use coal to displace more expensive fuels

i wonder what is more expensive that cheap large language models can displace? is the problem with selling unreliable B.S. really its price?

I think it's entirely possible that one categorically can't think correctly about markets and equity valuations since they are vibes-based. Post hoc, sure, but not ahead of time.

> 1) It's more effort and time than most people are willing to deal with, when instead they can create an account on Bluesky in seconds, and immediately begin sharing their opinions.

This is an interesting point because the apps and architecture behind Bluesky/Twitter/Reddit etc all have facilities for uploading, formatting and storing text, images, video - everything necessary for a website. So: why isn't there an equivalent mass-market "social website" platform that lets you create a website in seconds?

What stops that from happening - lack of demand? Are websites just not very useful for most people to have?

I think their top-down approach is a problem. What they call human civilization wasn't and isn't centrally-planned, and its goals and ideologies are neither universal nor implicit. The integration of software agents (I refuse to call them "AI") into civilization won't occur in a de facto cooperative framework where such agents are permitted to fraternize and self-modify. Perhaps that will happen in walled gardens where general-purpose automatons can collectively 'plan' activities to maximize efficiency, but in our broader human world, any such collaboration is going to have to occur from the bottom-up and for the initial benefit of the agents' owners.

This kind of research needs to take place in an adversarial environment. There might be something interesting to learn from studying the (lack of?) emergence of collaboration there.

Douglas Adams was satirizing touch interfaces and technological progress 45 years ago:

> A loud clatter of gunk music flooded through the Heart of Gold cabin as Zaphod searched the sub-etha radio wave bands for news of himself. The machine was rather difficult to operate. For years radios had been operated by means of pressing buttons and turning dials; then as the technology became more sophisticated the controls were made touch-sensitive—you merely had to brush the panels with your fingers; now all you had to do was wave your hand in the general direction of the components and hope. It saved a lot of muscular expenditure, of course, but meant that you had to sit infuriatingly still if you wanted to keep listening to the same program.

> quality really does not matter

What an inspiring vision for the future of news and entertainment.

go look at a photo of a CPU and imagine what else you'd need to run Windows on it, and then imagine it's probably extremely exponentially more difficult

it looks like the original submission is now unflagged and appears with a submission time of 1 hour ago, instead of 21 hours ago https://news.ycombinator.com/item?id=40725329

The weight of evidence is against Andreessen's ilk being able to improve the world because what have they done since they got wealthy? It's funny how the garage startup kids' mantra of "lean and scrappy beats behemoth" has changed into "the behemoth shall do as it may" at the exact same rate as their wealth has grown.

Lift with your knees. This is very difficult around change tables, cribs and baths, but very necessary, and nobody ever tells you!

Down in Singapore and Australia but not in Europe or North America, from what I can see.

As to how the FBI got access to the messages, Vice says[1] after Vincent Ramos of Phantom Secure was arrested in 2018, a confidential human source offered Anom, which the source was developing, to the FBI (probably in exchange for immunity or a reduced sentence, in my opinion). The source then seeded Anom phones to his existing distributors as a replacement for Phantom Secure phones, and from their they made their way into criminal organisations.

[1] https://www.vice.com/en/article/akgkwj/operation-trojan-shie...

Same with BitTorrent: first released in July 2001

Also, to expand on this point:

> My friends who ”made” it posted three quality videos a week. For 1-2 years.

For every ten people who did this, maybe one made it big. There's no guarantee the grind will pay off.

Cult! Ha!

Patreon makes more sense as a medium-sized business than an epic-scale corporation. Unfortunately reliable recurring revenue is a tough pill for "To The Moon!" investors to swallow.

But I thought Patreon never took more than a 12% cut...

communism baby

3% is nothing if shared equally

As early as 2012 the RAAF Hornet fleet flight activities (12-13000 hours per year) were estimated to exhaust the fleets' lifetime airframe limits (6000 hours) in 2020.[1] Keeping them in the air beyond this year was going to get more and more expensive and consume maintenance resources (and pilot training) that could otherwise be directed to the Super Hornets and F-35s.

Frankly I'm kind of glad someone took them off our hands before they became even more of a liability. We've got enough on our plate dealing with the F-35's corrosion issues (most of the fleet will be kept in moist salt air at Williamtown and be plugged into big dehumidifiers when not flying) and other things like its limited range.[2]

Which, now I put it that way, makes me wonder if maybe we should have kept a few Hornets.

[1] "43. The F/A-18A/B Hornet was designed for a safe life of 6000 airframe hours. At the current fleet flying rate of 13 000 hours per year, reducing to 12 000 from 2013–14, there is capacity on that basis for the Hornet fleet to continue flying until the end of 2020." https://www.anao.gov.au/work/performance-audit/management-au...

[2]https://www.aspistrategist.org.au/projecting-power-with-the-...

So, we ingest a third-party dataset that changes daily. One of our problems is that we need to retrospectively measure arbitrary metrics (how many X had condition Y on days 1 through 180 of the current year?). Imagine the external data like this:

UUID,CategoryA,CategoryACount,CategoryB,CategoryBCount,BooleanC,BooleanD...etc

When we ingest a new UUID, we add a column "START_DATE" which is the first date the UUID's metrics were valid. When any of the metric counts changes, we add "END_DATE" to the row and add a new row for that UUID with an updated START_DATE.

It works, but it sucks to analyse because you have to partition the database by the days each row was valid and do your aggregations on those partitions. And it sucks to get a snapshot of how a dataset looked on a particular day. It would be much easier if we could just access the daily diffs, which seems like a task Dolt would accomplish.

I mean it has a better chance of working than getting the third party to implement versioning on their data feed.

This page is a subset of NASA's broader ebooks collection: https://www.nasa.gov/connect/ebooks/index.html

People are using bitcoin to pay for cash, and vice versa. Besides drugs, that's its primary purpose.