gnfargbl

I feel like I did re-read the same article, or at least a version of it. Is this how content marketing operates these days? Find a popular blog post, rewrite it and attach your company's name to the end?

Previous discussion on an extremely similar article: https://news.ycombinator.com/item?id=49574167. This one even name-checks The Ironies of Automation in pretty much the same way.

Immediately and totally exiting the domain is about as close to admit you messed up and call it a day as you'll ever get from a modern company.

The statement can use whatever weasel words it likes, but the action pretty clearly says "we aren't mature enough as an organisation to handle this responsibility." Which would be quite the concern for me, if I were a customer for one of eMudhra's other product lines.

One repeated failure that I have noticed as I age is that is folks expect change to either happen quickly, or to not happen at all. As a species we collectively seem to lack the patience that reality demands of us, probably because our own useful lifespans are so short.

Self-driving trucks are not here, nor are they coming anytime very soon, but driver assistance systems really have advanced meaningfully since 2012. They will continue to do so, to the point where I imagine driving without them will feel pretty alien. I feel much the same way about AI; it will change everything, but we need to be thinking of that change on thirty or fifty year timescales. 2030 won't feel so different to 2025.

You could hack critical infrastructure before AI. Any and all of the bulk internet scanners have had lists of exposed critical infrastructure for quite a while now. At first it was shocking that nothing ever got done about it, then it became routine.

All that AI has done is to lower the bar of entry for criminal activity. Which is a concern, but it's not the primary concern. The primary concern remains that so much critical infrastructure is poorly secured.

The saying usually refers to emotional load, and not physical load. Which is very apt here: datacenter water usage is an emotive issue, but one which is completely meaningless in actual environmental terms.

An argument about factors of 10^-6 of a renewable resource, is not a serious argument.

We have real environmental problems, like excess CO2 in the air and forever-chemicals in our biosphere. The narrative must stay with those material issues and not be diverted by trivia.

Sure, and they'll be cognizant of that risk when considering expansion. The lessons of the dotcom fiber boom aren't that distant in memory.

Your argument in the post up-thread appears to be that because the Native Americans did not cultivate or make use of the land in any Western sense, they have no claim to ownership. The Kant quote attempts to repudiate this by stating that it makes no difference whether we consider the Native Americans' use of the land to be ineffective: their deprivation of it is still an injustice and, when it comes to injustice, the ends cannot justify the means.

I don't find Kant's position at all convincing, mostly because the two parties' understandings of the concept of "ownership" were not mutually intelligible (possession vs usufruction). However, the quote does have relevance to your post.

David J Nutt, Leslie A King, Lawrence D Phillips, Drug harms in the UK: a multicriteria decision analysis, The Lancet, 2010.

https://bigfatgenius.com/2220%20Fall%202010/Nutt%20King%20Ph...

Cocaine (in crack form) was found to be the third most harmful recreational drug overall, after alcohol and heroin. It also caused the most direct harm to users out of all drugs. Cannabis, not so much, mid-table. Non-crack cocaine was somewhere between the two.

Note that this study is now over 15 years old.

With respect, I think you're still missing my point, which is simply that bias introduces vulnerability. Without a real scenario to analyze, the scale doesn't really factor into the argument. A vulnerability is a vulnerability until proven mitigated.

> a tiny fractional loss in an RNG is not going to make a practical difference

I'm not so sure this is true. I don't think either of us is in a place to say whether this vulnerability has practical applications or not. A 1/65536 bias might seem like nothing important to you. It seems like potentially something to me, in a world where the attacker might control the volume of data generated.

We're talking about whether a modification of the expected probabilities changes the dynamics of the game. The example I gave was deliberately extreme, because that makes it easier to reason about.

If you want a casino example, then consider a roulette wheel that always lands on 36 but still pays out as usual. I think you'd want to play on it. Now consider one that always lands somewhere between 30 and 36. Still worth it, right? With careful bets and a good starting float you're still coming away from the table up (with a very high probability).

The point I'm making is that bias is exploitable.

It is absolutely untrue that a biased RNG has "zero practical impact." Modern cryptography has plenty of examples of relatively small biases leading to breaks. Check out Bleichenbacher's attack, for instance.

You could be correct that the very small bias here is not enough to be exploitable. But, given the history around this, it would be wrong to handwave it away as trivial.

Consider an 8-bit RNG.

By your argument, it would not be a problem if the RNG never generated 0. So, it must follow that it would also not be a problem if it never generated {1, 2, 3, ..., 253}.

That means that our RNG now only generates the values 254 and 255. Which of the values is generated is unpredictable on any given call. However, 7 of the 8 output bits are now always fixed and so completely predictable. Can you imagine how an attacker could exploit that?

Failing to generate only the number 0 is a weaker version of the same class of flaw.

> that implication is in the heads of people who believe AI is magic, or believe AI is advertised as magic, or believe that humans are dumb.

Yes! I believe that class of people are called "investors."

Your point is basically correct, I think, but the illusion of value that a corporate entity offers, and the reality of the value that the corporate entity is creating, are distinct things.

Most contemporary stories around AI include the implication that AI did something humans couldn't. This is because the big players have been shilling AGI hard for a while, and their valuations depend on maintaining the sentiment that serious progress in that direction is being made.

Which is a shame, because (to me) AI is already impressive enough without the oversell.

They didn't; one of the participants, Symon, is mixed-race (black father, white mother).

Does "without AI" mean without posts about AI, or without posts written by AI?

I think this is the former, but I want the latter.

In this domain, an apparent single unique piece of work is often composed of several breakthroughs. For example, when Andrew Wiles proved Fermat's Last Theorem, he had to develop multiple new pieces of mathematical technology to get there.

The claim here seems to be that the human mathematicians, working with AI, developed technology to go A->B->C. By training on those conversations, OpenAI was then able to encourage the model to go A->B->C->D.

In my opinion that situation should be acceptable, if openly disclosed, because it is in the public interest to make progress on these problems and because AI is clearly an amazing tool for making progress. But the human mathematicians are saying that OpenAI is presenting as if the model got from A->D entirely independently, without acknowledging their background contributions.

That link is a helpful contribution to this discussion.

I'm not at all familiar with this area, but my reading is that he appears to call it out as a relatively obvious extension of his own work:

> It is a creative and at the same time elementary construction that uses not just property (T) for an application of my result with Kun, but also for the ambient group G in order to overcome the problem, that the Γ-components might be of different size. Once this is achieved, the rest of the argument is straightforward.

Creative and at the same time elementary is where LLMs excel, generally speaking. It's why they are so good at writing code.

Grossmann collaborated with Einstein on GR, supplying quite a bit of the mathematical capacity required (which initially didn't come easily to Einstein). They published jointly, until Einstein was competent enough to work independently [1]. That's not equivalent to the situation being claimed here.

[1] https://arxiv.org/pdf/1312.4068

This is not a universal, however. In the UK, two-thirds of adults have a credit card.

> users have a private relationship with Google where they elect to let Google track them and retain their information

Elect to is a bit strong here. Acquiesce to, maybe, or perhaps suffer to.

One argument for LLMs is that although all information on topics X, Y and Z was already available somewhere, LLMs make that information more exploitable through collation, filtering and dynamic tailoring.

For a relatively narrow subject area (e.g. construction of pipe bombs) the collation is minimal, and so the filtering and tailoring probably isn't that important; a novice doesn't learn a lot more from the LLM than they would have done from a few Google searches.

For a broad subject (practical creation and exploitation of software vulnerabilities), the collation is very significant and the filtering means that LLMs can empower a novice to act at a similar level as an expert.

I would say that Bainbridge supports my position, if anything. Take a look at section 2.3 of her paper, where she lists all the reasons why simulation can be unsuitable in preparing for complex incidents: https://davidjusth.com/s/Ironies-of-Automation_Bainbridge_19...

I read the article before commenting, as is my habit. A disagreement is not always a failure to process.

To be very explicit, my problem is that the article argues that simulation will be a suitable substitute in the software space, as it is in the aviation space. I see no evidence for that; it's pure hypothesis taken as fact.

This article leads up to a hypothesis presented without evidence.

Yes, it's clear that if AI agents handle routine issues then only the most complex issues will be handled by humans. In no way does it follow that humans will be less prepared to handle those incidents than they are now; by definition, this class of problems have no rote solutions. Each issue of this type already requires deep system knowledge to remediate, and still will. Software is not aviation.

What we are seeing here is an instance of the more general trend where experts are still required to operate complex systems, but AI is destroying the career path that creates those experts. That's what we should be worrying about, not that people are going to be spending less time rebooting boxes. This isn't something you can fix with a "simulator."

...unless you're based in a jurisdiction like the UK, where an LLC may not get the same tax treatment as a C-corp (Ltd companies being the norm here).

No. Treaty of Paris, 1783.

Any plans for the n-hop version, ideally with a limit on the dwell time between successive hops?

I'd be interested to know how far I could get with, say, no more than two changes and no more than 30 minutes waiting at each stop.