iEchoic

You can still "win" by taking money from the other human players and minimizing EV loss against bots.

The major poker sites claim that they have really good (and very top secret) bot detection. I'm skeptical.

The "standard" strategy is to play GTO (game theory optimal). There are solvers out there (like GTO Wizard) that show you the "optimal" play for every situation, which is used as a baseline, and then players deviate to exploit specific player tendencies.

GTO trees are far too complex to fully memorize, so nobody can play perfect GTO. But you can do a lot of solver work to get reasonably close.

Four times a day, I get an email notification that someone requested a password reset for my Microsoft account, which gives me a six-digit number to recover my account. So every day, an attacker has four shots in 1,000,000 of stealing my account by just guessing the number. They've been doing this for years.

If the attacker's doing this to thousands of accounts - which I'm sure they are - they're going to be stealing accounts for free just by guessing.

I wrote up a security report and submitted it and they said that I hadn't sufficiently mathematically demonstrated that this is a security vulnerability. So your only option is to get spammed and hope your account doesn't get stolen, I guess.

Last time we were in Japan, we were walking around in Hakone and thinking the same thing. Then we stumbled upon a (literal) video game treasure chest. Inside there was a journal, stamps, and stickers to contribute your stories to.

https://i.imgur.com/VIPkCZr.png

Wish I took better pics, but here it was in the middle of nowhere. Japan is awesome.

Derek Lowe is a great writer. Worth reading his "Things I Won't Work With" articles if you want to read more of his writing.

https://www.science.org/content/blog-post/things-i-won-t-wor...

There are solo founders every batch, so realistic enough.

Very cool, thank you for sharing.

Has anyone done similar experimentation and/or benchmarking on using webgpu for neural nets in JS?

Making less money isn't really the risky part about founding a startup. The risky part is missing out on years of other life experiences, stressing (or losing) your closest personal relationships, failing and feeling personally responsible for disappointing everyone you convinced to believe in you, and developing an anxiety disorder (or worse) from chronic long-term stress.

Author's suggestion that they could have taken a "similar level of risk" as an early employee by taking secondaries as a founder is way off, IME.

Yeah, I was just curious to see if that's a purely modern phenomenon or if there's a longer historical trend.

It's unfortunate that subjective well-being was left out of this, because that's the one that looks like it might be getting worse. Arguably, it's also the most important.

You can definitely tell while you're playing it. It's probably the best-looking game I've ever seen and it has a pretty unbelievable amount of content.

FF7 Remake and Rebirth are both incredible, but Rebirth's sales are significantly lower. I'm hoping this doesn't make them reduce investment for the third installment.

Our lawyers told us otherwise.

Regardless of the answer here, the fact that there's still a debate about what basic functionality requires a cookie banner is really a testament to how bad this legislation is. How long has this been around, 20 years? And there's still widespread debate and lack of understanding as to what specific functionality requires a cookie banner?

> You don’t need a banner for the data that is necessary for the service to work at minimum level.

We were advised by our lawyers (a top SV tech law firm) that we should include a cookie banner in the EU even if we're only using cookies for functions like login. After eventually switching legal counsel (for unrelated reasons), we were told the same thing by our new counsel.

Either EU law covers cookie banners that use cookies for routine functionality, or it's so (deliberately) vague that even top tech law firms would rather everyone add a cookie banner than risk running afoul of the law. Either case validates PG's argument here.

I think this applies not only to startup founders but also to early-stage engineering teams. We distilled this into our company's engineering values:

Real work is work that improves our product or makes our users happy. This includes writing code, helping others write code, and helping hire people that write code.

#fakework is work that seems like work, but is mostly just vanity. Examples of #fakework include thought leadership, migrating things to shinier new technologies, and going to conferences. #fakework can be fun (and fun is worthwhile, too), but it doesn't count towards anything here.

Really interesting, thanks for writing this up. I'd love to see this applied to actually playing the game, provided that you could fit a (long) game state in the context window.

> As it was pointed out, no, EU regulations don't mandate cookie banners. It seem you have an axe to grind with the EU.

EU regulations do require cookie banners.

1999, created a Magic: the Gathering fansite on Geocities. Pretty sure I made a solid $13 or so with display ads, and would one day go on to make many dozens of dollars with my next Age of Empires II strategy fan site.

Good times.

> Great achievements have become team sports. You need multiple experts in disparate skillsets to achieve a huge business or great discovery.

The best way to get great people to work with you is to build things that people care about. That's something you can do alone, and it's never been more accessible than it is today.

Absolutely brilliant game.

GPT-4 is a major leap ahead of everything else I've used (including GPT-3.5), so definitely worth trying for comparison.

> But suppose AI is “only” a normal transformative technology, no more important than electricity, automobiles, or computers.

Most people who think the AI "race" is important aren't supposing this. So I don't think you can just "suppose" this away in the second sentence.

I use Todoist (todo list software). The killer features for me are that it has natural language parsing for recurring things like "take out the garbage every tuesday", "replace the A/C filters every six months", etc., and that it has mobile and desktop apps with quick-add hotkeys.

Basically, I can create tasks on any schedule or recurring cadence in 2-5 seconds by typing a sentence on any platform.

10b5-1 rules actually did change recently, IIRC it's now about six months from change to sell.

Check out the Madoff documentary on Netflix, there are interviews with people who had exactly that happen as they clawed back funds to try to make everyone whole.

Taking 60 questions across a handful of arbitrary categories and weighting them all equally is not a very useful methodology for this type of thing. It can help you differentiate Canada from Azerbaijan, but isn't going to be useful for comparing similarly-situated countries.

The OST is a masterpiece, and has only gotten better with time.

The Orchestra world tour (https://ffvii-remakeconcerts.com/) is incredible. Definitely recommend seeing it live if you get a chance.

Final Fantasy VII is still awe-inspiring for me. The original is arguably the greatest game of all time, and the remake is probably the best video game remake of all time, too.

If anyone wants to relive FFVII on YouTube, I'd highly recommend this[1] speedrun from GDQ. The runner beats it in ~8 hours and the commentary takes you through the story as it happens, sharing a lot of really cool insights and and highlighting details that I never knew from my playthroughs.

[1] https://youtu.be/EWgb-Jrk5tQ?t=677

> When a technology forum conflates a brand and a technology, one can feel confident their society is a clueless joke.

This technology forum would benefit from recognizing how out-of-touch and comically alarmist this statement is. Most of the discussions about Facebook here are starting to sound like some bizarre form of group hysteria.

> Do you use it more than hardcore fact analysis for decision making?

As it relates to this community, one area where I think quantitative analysis can be extremely limiting is in project management. I've seen processes where PMs take dozens or hundreds of tasks, subjectively evaluate a ton of values ("cost", "urgency", "importance", "impact"...) for each, and then use those to "calculate" prioritization.

In reality each task has at least thousands of deeply inter-related strategic and interpersonal dimensions, so any view we can quantify will be myopic. Other important dimensions include: "this task is a nightmare and the only person who can take it is Bob, who is key talent that is burnt out right now and might quit", "item A will greatly decrease in cost if we wait a few weeks, but only if we assign out item B to Person X because it's blocking and they can do it five times faster.", and "this task will help develop skills that I think are likely to be strategically important in six months, but only if we assign other items that leverage that strategic advantage".

Every variation on these that you can imagine is a dimension, and some of them are far more important than anything that you can easily quantify.

The human brain isn't good at this form of high-dimensional mathematical analysis. It is, however, extremely good at pattern-matching this intuitively. I have yet to see a project management process that performs better than someone with good intuition sitting down and deciding on gutfeel in 1/10th the time.