tw1010

Yeah I don't buy this idea at all. People freaked out over the ability to edit photos in dark rooms decades ago. Trust didn't deplete because pictures got tied to identities which in turn is tied to reputation (or by verifying "hey is that picture of the president k-wording a person with a gun real" with a trusted institution or something), etc.

There's other cool stuff that will come of this but the fact that this is the idea mentioned every single time shows a severe lack of imagination in our sphere.

People generally seem to have difficulty thinking about the diversity of types of lives on a population-level scale.

I have such trouble trying to figure out which of these many algorithms that are released will end up having a significant impact on a 5 year horizon. But this I have a rare hunch about that it could be quite significant (at least the direction in which it's trying to push).

Thank you

Curious: why isn't all porn sites sued for this? I don't quite get why they're only going after PH? Is it because it's harder to win a case against 1000 websites? Is it because it's more overhead to go after multiple sites and they expect the biggest return on investment only going after the biggest one? Anyone smarter than me who has a take?

I used to be sceptical that the concept of overtraining was a thing, but it definitely is. Just saying that running every day consistently might not necessarily be good for you in the long run. Listen to your body.

Pet peeve: when people take phrases intended for one thing and uses them for a slightly modified purpose. I take Branson's "screw it, let's do it" mantra as something entreprenurial, a call to break the rules (and stuff like that). Using it in this purpose feels more like "screw it, let's work even though I'd rather stay in bed", which is different emotionally, and kind of saturates the sayings power.

(Maybe nikes "Just do it" would fit better, but I get that its associations to big corporations isn't as well suited for the startup atmosphere we're surrounded by.)

I'm hearing this advice so often that it's pissing me off.

The book is BS. (I've read it.) There, I said it. It's always "this book is hugely thought provoking" (pointing at you Daniel Gross), and never ever and expansion on why or what insights it actually contains that's interesting. It has mildly interesting sentences that feels deep (mostly because they're confusing). The book has developed into some BS signalling device like Infinite Jest used to be. Everyone has read it, no one understands it. Everyone goes "oh yes, that's such a deep book, nothing has changed my mind like it since sapiens", and then we're all supposed to go silent to independently ponder it's many layered-ness, but in reality that's just what we do because we wouldn't come up anything remotely insightful if pushed into a corner. Frankly, the fact that this book is pushed so much makes me totally reconsider oft-repeated meme that "tech is low virtue signalling" (or low corruption). Clearly not.

(There, rant over. I'm overplaying how mad I actually am, I just feel like we need a few more rants against this book strewn about whenever this book is mentioned. Please, anyone, prove me wrong and a horrific narrow-minded dimwit by writing something more in-depth about what you think it contains and how it's insightful, I would love you infinitely.)

Like I said above, "typos" in sequences of calculations are obviously problematic, and lead (almost always) to mistakes in the final result. In this case that's not applicable since there's no "second arithmetic manipulation" following the typo:ed one. (The author replaces the incorrect one with the correct bayesian equation in the next section.)

There's value in collaboratively discussing half-baked ideas in public before any real applications have been produced.

Imagine this article as just another in a stream of posts trying to "think aloud" about BNNs, without any immediate pressure of applications for it.

I didn't say it was one of the most common mistakes, I said it's one of the most common mistakes I have observed. Purely subjective.

Not really, it's one of the most common mistakes I see in bayesian calculations. If the author was basing a lengthy series of calculations on that first step, it would be worse (but in this case the expression is quickly replaced by a corrected version for the classification discussion).

It's not "not getting the basics right", it's a simple typo. A thing like that shouldn't invalidate a whole article (unless you have skin in the game for the opponent argument).

Why does HN have a pattern of dismissing whole articles due to simple typos? It's as if we're so habituated to skim and do tldr-reading that our brain is working overdrive to find the slightest excuse not to have to do any type of reading beyond surface level.

I mean the boring answer is that people have patience and energy constraints. They arbitrarily ignore the whole group because VC isn't adjacent enough to their life goals, or something, to warrant the energy necessary to pick out the patience-stealers from the insight-givers.

They're not all bad, far from it. But sometimes a few actors poisons the well of a whole label (ehm, shark tank), causing some people to have an allergic reaction to the whole class (even if it isn't true that all of them are bad).

Isn't everything politics in some form or another?

It wouldn't be so sleazy if the fact that that's the transaction being made was less of an "open secret" and more of an acknowledged fact.

Curious: Do people find value in reading increment? I've tried reading it a few times but have often found it to be too many words for too little deep insight. Would love to be convinced otherwise though.

I agree with you. I find Melvyn Braggs rudeness quite useful. (And it doesn't seem like the quests mind. I mean, they're adults, they can handle some pressure to be succinct.)

Tbh, some anti-tech ideas (essentially eloquently presenting steelmanned ideas I'd only had at the periphery of my mind). All their ideas which goes against the grain, that I find valuable. (Warning tho: there's a lot of dopey-ness, amateurishness, and ideas in the podcast that I don't agree with. It's not for everyone. But I've found a lot of value in it the last month of listening to it.)

Eric Weinstein's The Portal really is amazing. A top 5 highlight of 2019 as far as I'm concerned.

Can't think of a specific episode, but all the controversial ideas discussed in the Red Scare podcast have really twisted around my brain (and put words to things I thought only I were alone in observing) recently.

Learning to become more creative so I don't have to chase new languages or frameworks for intellectual stimulation.

I feel like we're ignoring the obvious answer: because most people in AI right now are overwhelmed by the necessary mathematical background and comfort necessary to produce first-principles generative innovative algorithms and models, so all they're left with is following the latest trend, which is locked on deep learning right now.

Still so crazy that our field is so young that people like this are only now passing away.

The alternative is "radio" which is even weirder

Curious: Do people find value in reading increment? I've tried reading it a few times but have often found it to be too many words for too little deep insight. Would love to be convinced otherwise though.

The "coding is the same as writing" meme really should stop being subscribed to. It's soo different.

Some might hold you in higher regard, but probably that won't help you much in the interview process. Raw skill is what you'll lean on most of the work day, that's what they'll primarily judge you on. If they have two potential hires with equal skill but one has that on her resume, they'll go with the safer one. (I like Susan, I'm just trying to reason if being a whistleblower truly is benefit or not.) Most (thriving) employers are raw capitalists by selection pressure. Taking on extra risk by letting red flags pass through is not something top teir startups can afford.

I've never heard someone persuasively and eloquently express to me, in a non-dopey way, just how (small amounts of) acid can act as cognitive enchantment. Anyone care to take a stab at it?

For instance, is the effect only temporary, in the sense that it unlocks some region of not normally traversed mental state space? (E.g. allowing you to write paragraphs you wouldn't otherwise be able to.) Or is the effect more permanent, in the sense of teaching you something, providing you with a new muscle which can be used later even in a sober state of mind (teaching you to ride a type of brain bike, which then can't be unlearned), or by showing you connections between disparate ideas you hadn't previously had the ability to see next to each other?