As a prior I don't think it's at the level of "when you see a hack you should assume it's north korea"
443 karma · joined November 16, 2018
- The death of hundreds of thousands(?) in the ukraine/russia conflict
- Tens of thousands in palestine
- Tens of thousands in iran
They all seem more acute/no excuses are being made.
TBH I think I (and maybe others) just get triggered when the bogey man is invoked each time there's a crypto story.
1. Anthropic are probably losing money on the SpaceX deal or not making much at all.
2. To be willing to pay a spot rate so high anthropic likely have no better options.
3. Many of the existing cheaper vendors are likely doing "we will give you cheap compute for now if you give us big commitments"... (cf. hyperscaler capex spends)
Which I suppose is to say that if the spot prices are in the same ballpark as future costs then anthro stops looking profitable pretty quickly. If you're especially bearish you can also throw in the huge reduction in token prices recently and that SpaceX is charging google almost 2x as much as they charge anthropic!
Happy to be corrected by someone more knowledgeable.
https://www.bloomberg.com/opinion/articles/2019-06-26/everyt...
He has decent business takes on tech, the podcast/articles are worth paying attention to. The actual tech takes are probably too dumb for the hn audience, but you aren't the target audience.
1. 5900 was always known to be an extra unsafe port to have open
2. all ports should be assumed unsafe without explicit vetting
I'm assuming the argument being made is (2), but it really is a sad safe of affairs because use cases from time to time do involve having ports hot.
Or should things be taken one step further to *always* use a vpn or similar?
Offsite backups are sadly rare these days, and aws sales is the main reason why.
Then one day when I was 18 I went to the pub and played pool with/chatted up some girls for one of the first times. When I got home I went to bed and closed my eyes and could see the whole scene vividly for the first time ever. I had been drunk plenty of times before so assumed I had been drugged, told my parents and just got a nonchalant 'wtf' reaction.
Since then I've mostly assumed it's a skill issue, and in the past 15 years have maybe visualized ~3 times under similar circumstances (usually sleepy, usually sexual).
I've spent a few meditative sessions trying to train, and gotten the sense that it might be close, but it seems like it would take a good chunk of dedication on my part to make it happen.
Modern politeness encourages us to think of all humans as more or less the same, which I think causes a default minimization of difference mindset.
One thing that really blew my mind in my 20's was learning that asians have a different type of earwax -> usually white and dry instead of yellow and waxy. It really shocked me that I was never taught that in school, and if people are unaware of common physical differences I am sure we are largely unaware of common mental difference.
There's just a substantial bias where people assume all 'new knowledge' is wrong. Probably correct most the time, but still often wrong.
I spent about a year as a consultant in the AWS space, visited about ~15 clients of varying sizes.
More often than not there's a single pg aurora instance responsible for 50%+ of the bill. Even worse are the serverless aurora offenders.
All the indexes and guarantees of PG don't come cheaply and dynamodb pricing is not cheap but comparatively reasonable. It really is a good product if you know how to use it.
You think you are good at chess because you win against your neighbour, then eventually a class mate absolutely embarrasses you multiple times and you realize there is a whole world you don't know, and before now you didn't know that world existed. Then you eventually beat your classmate before getting whomped by the beginner in the chess club and the whole cycle happens again. After this happens a few times, you should begin to find each time less surprising than the last.
To me this is the essence of DK and not really something to be proven/disproven by maths. You don't know what you don't know, and it takes some learning before you are wise enough to realize there are unknown unknowns.
> instead showed that both experts and novices underestimate and overestimate their skills with the same frequency. “It’s just that experts do that over a narrower range,” he wrote to me.
I think this is trying too hard on the researcher's part... if the data were reinterpreted such that small estimation errors are considered correct then the effect would reappear.
Do you have a basis for that?
It seems like in life some tasks reward overconfidence, e.g. starting a business, and some tasks reward underconfidence e.g. retirement savings.
I'm not sure there is a strong bias one way or the other.
I'll admit I don't fully understand what they were claiming/how it changed over time but they later issued an apology for the incident [1].
From what I can work out the thrust of the abc claim is "don't blame our journalists, the editors on multiple occasions accidentally did this for no reason at all and we then failed to catch it".
Then midway through the apology they pivot to blaming twitter, or at least that is the flow of the guardian article.
[1] https://www.theguardian.com/media/2024/sep/17/abc-australion...
A recent, prominent example is [1] where they edited footage to make it look like a soldier was shooting at an innocent civilian. They then appointed a 40 year abc/sbs veteran to investigate the incident who unsurprisingly found no wrong doing.
All of this on a topic that isn't even super political...
Granted they are still probably the best journalists in the country, but when you can't trust them to not alter primary evidence, or admit real fault when something has gone wrong, you cant really trust them for much.
For day to day stuff I still find them an ok source, it just needs to be taken with a very heavy dose of salt.
[1] https://7news.com.au/spotlight/7news-spotlight-abc-launches-...
The task was to write a simple sum of a list, but they had to convert the list to a pandas dataframe and then call some series of arcane functions. I originally thought this was just a preference but when I asked them for the simpler version while standing over their shoulder they were unable to do it.
They had ~8 years of claimed experience at the time and were from a country notorious for faked credentials. Turned out to be an ok hire overall (friendly, hardworking) but definitely lacking 80%+ of their supposed competencies.
Ah yes, the much revered cosmological fairness constraint.
> Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4)
- A way to convince others of your position.
- A way to demonstrate that you did research (because experience is worthless)
- A way to soften harsh feedback.
At first I found this very dysfunctional but now I don't mind too much. It beats being ignored and/or offending someone.
Intel's foundry costs are probably competitive with nvidia too - nvidia has too much opportunity cost if nothing else.
The slightly altered "explain when and why these techniques are *not* necessary" is much less appreciated.
The problem is that many of these trade-offs only applied to older databases. The more relevant axis is about how distributed the db is, the replication type etc.
"America arms Israel to attack Lebannon and Palestine"