HN user
While the reasoning behind this is dumb and it is unfortunate that the screening is mandatory for over 30s and not voluntary, I guess I can’t hate that some people will have the choice of receiving a treatment they may want.
Seems like just a matter of time until this logic extends to mandatory HGH and anabolic steroids for all soldiers or something.
45k/year is excellent! If you believe business insider, U of Chicago’ll run you 100k / year. https://www.businessinsider.com/most-expensive-colleges-tuit...
Traditionally the story was that almost no one paid the sticker price, but still that’s an eye watering sum even discounted.
I’m sure in the next few years we’ll have stories of people 500k in debt or more for their schooling.
Every time someone DIYs one of these guys the internet at large lambasts them as fire risk death traps. Fairly curious if these are any different, certainly looks like they’re much the same from the photos.
I mean, god willing, but it'll be just as likely that we'll blissfully consume 100 million token contexts in that case.
As they saying goes, you have to dodge the trap every time, they only have to get you once. Sooner or later we all will slip up if we’re subjected to endless con attempts.
Shame that the guardian describes Powell’s statement as ‘blistering’: https://www.federalreserve.gov/newsevents/speech/powell20260...
It reads to me more as ‘tepid’. He’s obviously trying to be as apolitical as possible, but I don’t think that’s going to help.
If Trump and Rubio are not credible, then there is no way to determine the intent of any military action, so the bet is impossible to evaluate.
That’s pretty funny.
Heh, it’s impossible for me to look at these without hearing the launch theme from Apollo 13 in my head. Such a glorious programme.
I think this sums it up pretty well.
"When I am weaker than you, I ask you for freedom because that is according to your principles; when I am stronger than you, I take away your freedom because that is according to my principles,"
I guess they can try, but in this era of ubiquitous surveillance it’s hard to imagine they would succeed in secretly entering a home.
Isn’t this more that they’re buying a part of an upstream supplier than a client?
I suppose the hope is that they don’t, and we wind up with commodity frontier models from multiple providers at market rates.
I guess these things aren’t literally exclusive, but it’s pretty amusing that elsewhere the rebuttal argues that we’re deep in a great stagnation which we need space exploration to pull us out of. (In the bit where he is arguing against the idea that we should wait a century and then maybe try to colonize space with greater technology)
The slowdown in GDP growth is not mere paranoia, but an economic fact. Part of the problem with seeing clearly the stagnation all around us is that we need to compare ourselves to what might have been, not to the 1950s as the authors do.
I was very interested to see in that rebuttal that they explicitly called out ‘datacenters in space’ as a means of ‘exporting’ solar power to the earth.
As the Weinersmiths point out, the ease of generating solar electricity in space is foundational to space development. They focus on the challenges in beaming power back to the Earth, but the “power” could be returned to the Earth in other ways, such as by doing energy intensive manufacturing in space, with the result that we do not need the power on the Earth itself. One modern idea that O’Neill did not consider is to move server farms in space, where power is cheap and you can dump heat into space with a black piece of metal. If this was done on a large scale, the carbon impact of data services on the Earth would drop greatly even if power is not beamed back to the Earth. There are almost certainly other ways we can use power in space to do things in space that benefit people on the Earth.
So the original article seems to think that cooling is a significant challenge and that solar power in space is not ‘that much’ more effective than on the earth, and the other that cooling is trivial and that solar power is easily obtained. I’m inclined to go with ‘space is hard’ as that seems to comport with my other readings, but obviously the critique of ‘a city on mars’ is advocating for space exploration and is so motivated to minimize the difficulties.
I slightly have trouble believing that Mr “Stop wasting tokens by saying please to LLMs” Altman is not considering how his models can be optimized. I suppose the real question is how accurate are the utilization numbers in the article.
Indeed. The moral of the story is that it’s ideal to do illegal acts which are difficult to undo, as it’ll give the judiciary greater pause.
Pertaining to that observation, I really liked this section:
In 2022, California became the first of a half dozen or so states to offer free school meals to all students, regardless of family income. Dillard supports free meals for all students with an emphatic, “Yes, yes, yes!” Food should not be based on income, she says: “It should be part of the school day. Your transportation is of no charge to students. School books are no charge to students. School lunch should be of no charge to students. … It’s just the right thing to do.”
On one hand, that seems like an excellent argument to use for free school lunches. On the other hand, it feels like school busses are like libraries, accidents of history out of step with the modern world. If this became a rallying cry there'd probably be a strong pushback to start charging kids to be taken to school.
Of course it doesn’t. But there is a difference between treating the story as an oral tradition and the explicit and unerring word of god.
The only advantage is that if the company is more efficient they'll be less likely to fire you because the business is failing. They'll just be firing you to eliminate a cost.
Did you ask it to search the Internet as a part of your request? It is still extremely imperfect, but that typically helps it get basic details correct. At least for me at any rate.
Man, it's wild how we have such different experiences. I have a base level M1 air, and I feel like it chugs along anytime I ask it to do anything even vaguely computationally expensive. Obviously that's not rigorous, but that's my subjective impression.
I’ll defend airbus a little - there are flight laws that more or less provide at any given moment as much automation as is possible given the state of the sensors and computers. So it doesn’t just go ‘oops, a sensor failed, now you have direct control of the plane.’
It does have the same problem - if 99.999% of your flight time is spent in normal law you are not especially ready to operate in one of the alternate laws or god forbid direct law, which is similar to the case of a driver who perhaps accustomed to the system forget how to drive.
But I think we have a ways before we get there. If the car could detect issues earlier and more gradually notify the driver that they need to take control, most every driver at present retains the knowledge of how to directly operate a car with non-navigational automation (abs as you mentioned, power stearing, etc)
A ‘cyber-related disruption’ is a great euphemism. I suppose all them sci-fi novels had cyber attacks happening like the weather, but it still feels very strange to live it.
It's a pretty rough headline, clearly the author had fun performing the test and constructing the thing.
I would be pretty regretful of just the first sentence in the article, though:
I ordered a set of 10 Compute Blades in April 2023 (two years ago), and they just arrived a few weeks ago.
That's rough.
Though I'm sure many will recoil in horror, it makes perfect sense. Modern language models are typically knowledgeable and non-judgemental, especially if you have a question which you fear reprisal from whomever else you might confide in or ask.
Given the concreteness of the output they provide, I'll be quite surprised if there aren't people out and out worshiping language models before the decade is out.
Presumably other search companies will be on a more even footing with google with regards to what they know about you to sell to advertisers if google doesn't also have your complete search history and a complete history of everything you do on your phone.
Hasn’t the US already done that? Huawei’s AI processors are banned in the US, with an effort made to ban them in US allies worldwide.
I wonder if this means the US is going to come for Global Foundaires, TI and Micron to extract an equity stake too. Interesting times.
Admittedly, just because something hasn’t bourn fruit doesn’t necessarily mean it’s a bad plan. All they need is a major stumble by TSMC or AMD or some tailwinds themselves and they could start looking much better.
Across client and data center, I’ve directed our teams to define next-generation product families with clean and simple architectures, better cost structures and simplified SKU stacks. In addition, I have instituted a policy where every major chip design is reviewed and approved by me before tape-out. This discipline will improve our execution and reduce development costs.
One would hope they would figure out if a design was a good fit for the company well before tapeout. Ideally before they were done with the rtl.