HNHacker News
TopNewBestAskShowJobs

imperfect_blue

143 karma · joined June 28, 2020

submissionscomments
imperfect_blue··on Show HN: A highly opinionated, fully functional Obsidian vault
If all you're using Obsidian as is a fancy markdown editor then of course you lose nothing. But the whole point of PKM apps is that they're more than a fancy markdown editor.

Without the tool, a Kanban board that remains as a bunch of scattered markdown files is worthless for its intended purpose, and not easily recreated in another app.

It's like, a Microsoft Word document is perfectly portable if you never use any of its formatting features and always save it as plaintext, but at that point is it really using Microsoft Word?

imperfect_blue··on Show HN: A highly opinionated, fully functional Obsidian vault
This is true for every note-taking app with an export functionality. And it really isn't good enough.

The entire reason you use Obsidian or Notion or Anytype or Capacities.io or Logseq, or anything more powerful than a set of markdown files in the first place, is that you care about the relationships between the notes and the rich functionality that comes with them - sorting, search, linking, reminders, etc. None of this is easily exportable into another app. The series of markdown files that make up Obsidian is not a substitute for the Obsidian software, nor are the relationships easily rebuilt in another app.

imperfect_blue··on Peter Norvig critically reviews AlphaCode’s code quality
To be more accurate, the 7 year old will tell you they can't do it and suggest the use of a calculator.

GPT simply gives you a complete hallucination. This can be gotten around imperfectly with prompt engineering to get GPT to admit it doesn't know. But still, hallucinations are pretty dangerous failure modes in many applications, and definitely something we don't want in a future iteration. I don't know how much of a change it is to create a GPT without hallucinations, but I suspect that it is non-trival.

imperfect_blue··on Peter Norvig critically reviews AlphaCode’s code quality
A 7 year old would know that you would want to use a calculator, and indeed, would be capable of multiplying those numbers with a calculator.
imperfect_blue··on Twitter has re-suspended ElonJet account
Wait, where did he say this?

To me it doesn't just feel like extortion, it straight up is extortion.

imperfect_blue··on Fosshost Is Shutting Down
The purpose of reporting crime is not to receive compensation; it's to create a norm where crime is punished and hence deterred. It's a societal benefit not a personal one.

It's not good that miscarriages of justice do happen, but that doesn't change the overall point.

imperfect_blue··on Estonia saved its citizens 820 years by moving public services online
It's almost a tautology. All things being equal, the less time spent on bureaucratic processes, the better.

Its just so that governments (and academic institutions, nonprofits, healthcare, etc.) are the natural environments for bureaucratic bloat.

imperfect_blue··on AI-generated art sparks backlash from Japan’s anime community
I am also a thief in all of the same ways.

When I write, I am doing nothing but laundering sentences and grammar rules from my English textbooks as well as sampling from the millions of sentences that I have read in my lifetime to compose into this comment as a response to your prompt. When I code, I heavily reference existing implementations and often only make minor stylistic changes, copying algorithms and data structures wholesale from memory. When I cook, I'm literally copying the contents of the cookbooks or cooking channels with a few minor substitutions at best.

Unfortunately, I am not brilliant enough to come up with my own algorithms, recipes, or literary flourishes. Few people are truly "creators", to be honest, and even creators build off or derive from existing work.

imperfect_blue··on Science, technology and innovation isn’t addressing world’s most urgent problems
Also unlike a video game, the tech-tree dependencies are not obvious. We don't know in advance whether a tech will be a productive avenue of R&D, or if it depends on other unrelated advances in theory, material science, or computer modeling, or was actually completely unfeasible in the first place. I find wind power [1] as a case study representative, and we see similar stories for e.g. fusion, cancer research, AI

[1] https://constructionphysics.substack.com/p/why-did-we-wait-s...

imperfect_blue··on People Staring at Computers (2021)
I don't know, seems like a pretty proportionate response to someone coming into your store and installing what's effectively spyware on your machines.

Sure, no harm was done and it was in the name of art, but how would anyone making the decisions at Apple know that?

imperfect_blue··on As Covid hit, Washington officials traded stocks with exquisite timing
Insider trading improves the efficacy of financial markets actually, since it allows markets to take non-public information into account. The bigger problem is that it opens up avenues of market manipulation and causes conflicts of interest.

It's the same conflict of interests problem here. Are your politicians acting for the sake of the public, or do they have their own financial interests in mind when putting forth a policy? E.g. if they've made financial commitments in expectation of a policy that will benefit that investment, that can most certainly influence their decision on said policy.

imperfect_blue··on “When are we going to use this in our everyday life?” (2017)
You certainly don't need calculus to understand exponential growth; you'd get a good sense of it from just looking at a chart.

The ability to apply learned theory to real life in cross-domain ways is not common and not so easily taught by rote. Teaching calculus to everyone won't solve this.

imperfect_blue··on U.S. appeals court rejects big tech’s right to regulate online speech
In the hypothetical world, where press is a natural monopoly, we would be far less likely to allow the press to refuse to publish viewpoints they don't like. Or we would outright nationalize the press and lead to a different sort of problem.
imperfect_blue··on Stripe has decided to nuke my entire business
There's no need to reverse it even, just go through your standard expense claims procedure or product refund procedure.
imperfect_blue··on Stripe has decided to nuke my entire business
Debit revenue, credit employee expense claims, journal entry: adjustment for test purchases made. Pretty standard, doesn't affect your numbers at all, no need to make any note in your statement (which you wouldn't have to do anyway because it'll never meet the threshold for materiality). Its not some sort of difficult accounting problem.
imperfect_blue··on I will now send this to beginner programmers
Overscoping a beginner project is completely harmless. The worst thing that can happen is that you get stuck along the way and realize that a feature is much more difficult or out right impossible. You then cut that feature or give up on the project if you've sunk enough time into it.

Fortunately, actual delivery of a beginner project is pointless. The only point is to learn by doing and you accomplish that regardless of whether your project fails. And failing projects is a lesson in itself on scope, estimation, and research; you won't get this experience if someone else is scoping for you like in a school project .

imperfect_blue··on Dear startups: I hate subscription services (2021)
They do it not because they want to but because they have no choice. The UX on those software are horrible by modern standards and incur massive wastes of manhours.

Had a subscription model been used, those software might have benefited from 30 years worth of usability updates...or they might still be garbage, looking at SAP.

imperfect_blue··on I accidentally built a spying app
All of these safety measures are developed through trials and iterative improvements, not something that was foreseen by the inventors. That's because its difficult if not impossible to accurately predict the impact of a revolutionary product at the onset. Its hard enough predicting whether something will be revolutionary, much less its impact.

This iterative process is working fine for social media. The platforms are changing, how the users use them is changing, it just needs more time to mature.

imperfect_blue··on I accidentally built a spying app
You can run people over with cars, electrocute them with electricity. This doesn't mean we should never have built cars or generators.

The more powerful and impactful something is, the greater the potential for abuse. You don't even need that something to be abused - bad people benefited from electricity just as much as good people. This doesn't mean we should stop building powerful and impactful things, or think of ways to restrict them such that only the "good guys" can benefit.

imperfect_blue··on Ask HN: After Slate Star Codex, where are the nuanced discussions?
...then it's great that at least they're talking about it on some forum, giving us a chance to persuade them otherwise and to call the police. If we didn't have such a forum, they might already have started their murder spree.
imperfect_blue··on Ask HN: After Slate Star Codex, where are the nuanced discussions?
I interpreted the parent to that assume his hypothetical individual is advocating for the murder of some people for what they believes to be good reasons, since nobody short of the mentally ill will actually ask for random people to be killed for literally no reason.

I can list 10 people that the world would be better without — the likes of Kim Jong-un and Ayman al-Zawahiri. That's not too different from wishing for their death. Plausibly we can discuss whether it's a good idea for them to be assassinated, and whether the power vacuum would just cause another worse despot to replace them.

So it seems quite conceivable to me that a discussion could be had about the reasons, feasibility, ethics or geopolitical implications, or about why they believe that someone does or does not deserve killing.

imperfect_blue··on Ask HN: After Slate Star Codex, where are the nuanced discussions?
I reject that discussion will be inherently flawed without representation. I further submit that representation can cause flaws in discussion.

To use your thought experiment as an example, say if we are critically examining Sumerian religion and their creation myths, we may not want the Sumerians in the room even if they were readily available, since they may well take offense at us disrespecting and dissecting their beliefs. If we are discussing Sumerian religion with Sumerians and we don't want to cause them offense, we would have to tiptoe around the fact their gods well, aren't real, and subjects like "what might have inspired the Sumerians to ascribe this trait to that god" can't be discussed at all.

And to use a example more grounded in reality, a group of men can't have properly conduct a discussion on how to attract women with women in the discussion. I am well aware that this sort of discussion has pejorative connotations with how infamous the PUA community has become, but the fact is that young men do need to social spaces and groups where they can learn this, since it is skill that needs to be learned and practiced.

Or to use an even more absurd example, a group of rape victims can't have a discussion with a rapist, even though the rapist would be able to add their side of the story to the discussion.

imperfect_blue··on Ask HN: After Slate Star Codex, where are the nuanced discussions?
The original question was

> how much of this nuance is achieved by having discussions about issues which have real life-or-death impacts on people but without having any of those people inconveniently present?

Followed by an assertion that

> A group of X people is [not] capable of of empathising with, understanding and having a nuanced opinion of the concerns of !X people

So I think there are people advocating that you should be speaking only if you lived it.

Of course, anecdotes and personal experiences are valuable and including people with personal experience is helpful, but the central claim here is that you can have productive discussions about X without anyone with personal experience with X.

imperfect_blue··on Ask HN: After Slate Star Codex, where are the nuanced discussions?
To add, personal experience can be useful, but is not the end-all in decision making. It's useful in the ways anecdotes are useful, that is to say, good as an illustrative example but may not be statistically representative.

In general, you do not need to have personally experienced X to have a useful or "valid" opinion on X. As a reductio ad absurdum, nobody would be allowed to have an valid opinion on murder.

imperfect_blue··on Ask HN: After Slate Star Codex, where are the nuanced discussions?
But we can in fact have a conversation about whether we should murder 10 people for no reason, it might even make an interesting discussion on ethics and moral philosophy.

We can agree to conduct adversarial studies with flat-earthers and examine the evidence together.

All of this is possible as long as everyone is discussing in good faith and open to the possibility that they might be wrong. That should be the only requirement, rather than fixed boundaries for what's acceptable to discuss.

imperfect_blue··on Yann LeCun quits Twitter amid acrimonious exchanges on AI bias
> China might be using AI not to "round up" Uighurs, but as an intelligence measure to prevent more terror attacks.

That is exactly what I claimed they were potentially using it for.

I acknowledge that I am very much biased against China as I live in a country within it's sphere of influence and have friends in HK.

But in any discussion of using AI for unethical purposes, China is the ur-example, as Nazi Germany is to facism: an authoritarian government with a history of tech surveillance, censorship and media control, and minority oppression, and the tech to back it up.

If they don't like that, maybe they should stop using technology to oppress their own citizens.

imperfect_blue··on Yann LeCun quits Twitter amid acrimonious exchanges on AI bias
If Bagger 288 tries to drive across a footbridge that's not really the engineer's fault.

Although while everyone intuitively understands the load bearing capacity of a footbridge, not many understand the capabilities of ML models. So perhaps better advertising the capabilities of AI would help inform decisions.

On the other hand, nothing an engineer or scientist can do will stop the Chinese government from using their technology to predict and suppress dissidents and minorities.

imperfect_blue··on Yann LeCun quits Twitter amid acrimonious exchanges on AI bias
To clarify, The Tutorial on Fairness Accountability Transparency and Ethics in Computer Vision is not a technical presentation, it is very much a intersectionalistic take on the topic, more concerned with power, oppression, and social justice than engineering.

Suggested technical solutions [1 above] was exactly one sentence and the only sentence in the presentation that suggested any sort of technical approach, presented without elaboration, and the recommended solutions like Counterfactual Testing are also methods for fixing the training data.

The remaining 99% of the presentation was spent on the idea that if AI engineers and scientists were more diverse and attended more diversity training, the models they produce hopefully wouldn't be biased.

imperfect_blue··on Yann LeCun quits Twitter amid acrimonious exchanges on AI bias
I think the Yann's main point is that this specific model's racial bias was a training data problem; Gebru has not provided any evidence otherwise, as far as I know, apart from repeatedly asserting that better training data won't fix the problem.

It would indeed be a valuable contribution if Gebru could have shown how better training data would not have fixed the problem (e.g. by feeding better training data to the same model and still reproducing the issue).

But as it is, she is asserting that better training data won't help without proof and not recommending any alternatives, while taking a hostile tone of conversation. I would be hard pressed to take her claims as good faith criticism.

imperfect_blue··on Lemmy, an open-source federated Reddit alternative, gets funding for development
Early Reddit isn't what what most would call a "bad result". Ditto for early 4chan and for a lot of smaller communities that have very little restrictions on content beyond civility and anti-spamming.

The problem is that having uncensored speech doesn't scale, because eventually you'll become big enough to attract media scrutiny, which inevitably cherry-picks the worst parts of the user base. This cause 1) advertisers to pull out and 2) starts attracting more of the wrong type of users.

← PreviousPage 2 of 2