HN user

cmuguythrow

219 karma
Posts3
Comments53
View on HN

I would support efforts to try to redraw the line as stated above. I do think we would all be better off if we could do so. But my default assumption is that that won't happen - we will fail to organize effectively enough to redraw the line and we wont "get the day off" that the parent asks for. Maybe I need to be more optimistic to encourage such efforts, but because I think merely asking has such poor chances, I'd rather point to the problem of coordination in an attempt to move the convo towards "how do we beat the coordination problem" - hopefully by being more aware of the size of the hill we are facing, we can be more likely to devise more ambitious approaches to solving the problem

Hours alone is not the only thing that matters, but all else being equal it does matter. Someone who works smarter for 20 hours will get ahead of someone who works stupid for 40, but if both people work smart the 40 hours person gets more done

sadly life quality is not the thing that the competitive system is maximizing for, and thats one of the article's points. we compete to our own detriment, but to not compete is to become extinct

Good source for what? I'm just trying to point to a concept, an idea. There's no "facts" here, just speculation. If you disagree with the point there, why don't you just say what you think is true instead, I'm happy to discuss the ways in which the article is wrong

Relevant: https://slatestarcodex.com/2014/07/30/meditations-on-moloch/

the concern would be that this new ability will actually increase competition and give us less than we had before

this is not something that can just be blamed on the "CEOs/execs/shareholders" of the world. it is evolutionary competition - unless we can ALL join forces to draw the line somewhere, someone will choose to defect from the agreement to "just work less", because doing so will make them succeed at the expense of others. even if everyone from one country agrees, the other competing country that defects and works 996 with agents will "win" and conquer the lazy country.

I wish I knew what to do to fix it, doesn't seem sustainable but I don't know how to make all of humanity cooperate without doing something even worse

Site feedback:

- Can't view demo until dismissing cookie banner (which I didn't see since its like ShadCN or whatever like the rest of the site). Thought site was broken for 20 seconds

- Demo step 3 animation looks strange/broken

- Obvious GPT copywriting in site and in this post (turn-off)

Notes:

- I don't think I am your target customer. I run a seed stage startup and am the only sales person. I don't do things like MEDDIC/BANT, although maybe I should

- I use fireflies for sales calls (and all calls) currently - just to have a searchable transcript of the call. I would replace Fireflies if I thought I would get enough of a benefit by doing so, but I currently don't - I don't feel like I need the coaching. I wouldn't want two notetakers in my call

- How could I make sure the bot only joins/analyzes the calls that are sales calls? How does it know which stage of the funnel the person I'm talking to is? Would be worried all I get is slop without the context needed to filter things appropriately

Hope this helps!

I would be in favor of anything that improves the current political system, including a shot at this policy. On a meta-level, I would even be in favor of new political processes that are WORSE, simply because the adoption of such a policy could prove to people that we CAN change our processes, and then we could (try) to continue to amend our process until we find one that works.

My personal favorite approach at the national level would be Ranked Choice [1], as that would preserve the (IMO important) single decision maker in the executive branch, while removing the incentive to vote for someone you hate just because they aren't as bad as the Other Guy. Interested to hear if HN knows of other/better ways to accomplish the same

[1] https://en.wikipedia.org/wiki/Ranked-choice_voting_in_the_Un...

Magnetic | https://magnetictax.com | Multiple roles | San Francisco (ONSITE) | Full-time

Magnetic (YC S25) is building the AI tax preparer for accounting firms. We use vision models to extract and structure data from tax documents and automatically populate legacy tax software, saving CPA firms hours of tedious data entry per tax return. Firms are coming to us faster than we can build - our next tax season is already 10x the size of October's, and we have a waitlist for several X more.

We just raised a $5.1M seed round and are building a small, senior, in-person team. We believe in everyone talking directly to customers, and getting into the workflow themselves (i.e. learning how to do tax returns). We like to work together all day, whiteboard new ideas, get lunch together in person, and get drinks after work.

[1] Founding Engineer – https://www.ycombinator.com/companies/magnetic/jobs/77FvOwO-...

[2] Founding Tax Expert – https://www.ycombinator.com/companies/magnetic/jobs/eReV8hh-...

Email thomas @ website to apply

I have often wondered why the government doesn't do anything about this. Is the science not clear enough yet?

Quick search shows that we knew about lead hazards as early as the 1920s/1930s, but it took until the 1970s to get regulation about lead paint and gas - hoping we don't repeat that in this case

Air travel can be revolutionized with a vacuum-sealed passenger transport system, reducing total airport time to just 10 minutes while maximizing efficiency and safety. Upon arrival, passengers would undergo a rapid preparation process, where an oxygen-rich biofluid is introduced through both mouth and anus, ensuring full oxygenation without the need for active breathing. They would then be gently encased in a flexible, vacuum-sealed transport sheet, immobilizing them completely to eliminate security concerns and expedite boarding.

These passenger sheets would be automatically sorted and loaded onto aircraft via a conveyor system, optimizing cabin space with a precision-stacked configuration. Without the need for traditional seating, aisles, or security screening, aircraft capacity could increase significantly, lowering ticket prices while enhancing operational efficiency. Upon landing, passengers would be swiftly unloaded, revived from stasis, and ready to continue their journey within minutes.

This system would not only eliminate delays caused by baggage handling and TSA procedures but also enhance fuel efficiency by optimizing passenger distribution within the aircraft.

Yeah DuChinese is I think the premier Chinese graded reader app right now. We also have the tap-to-lookup and make-flashcard-from-content flows (unfortunately only in our system for now, haven't build flashcard import/export yet). The thing we have that they don't is the ability to generate content about whatever subject you want (which can help make it much more personally interesting)

Also while naming quality resources I should also mention [Pleco](https://www.pleco.com/) - it's _definitely_ the best Chinese dictionary app - highly recommended.

Yeah this is a huge problem. AFAICT the "solution" currently is basically to just grind to the point you can understand Peppa Pig or the rough equivalent at ~40%-50% comprehension, and then sentence mine each episode painstakingly using something like Migaku + Anki flashcards until you can watch a brand new Peppa episode at ~80% comprehension. Its painful but after this you really "unlock" content and its a lot smoother and more interesting after that (the stone finally starts rolling _downhill")

I'm making an app to try to help with low level comprehensible input, posted elsewhere in this thread

I’ve been learning Mandarin via Comprehensible Input (CI) for about 9 months and really admire OP’s dedication and consistency. In the first 4-5 months of being truly consistent with ~1hr a day of Anki and Peppa pig I got to around 2,000 words and was able to have a great experience when I traveled to Taiwan, so I can vouch for the core methodology in this post. It’s not “easy”, but it’s definitely the most effective way to learn a foreign language that I know of.

The CI community has come a long way in the last ~5 or so years - the general consensus looks a lot like OP’s methods, which I would summarize as:

1. Brute force [premade Anki flashcard decks](https://ankiweb.net/shared/info/810519009) for the first ~1k most common words

2. Start watching comprehensible input as soon as you can, ideally for an hour a day or more

3. [Sentence mine](https://www.youtube.com/watch?v=QBcQJESGQvc)the comprehensible input and add it to the daily SRS flashcard grind

The best summary of these methods that I’ve found is https://refold.la/

Self plug: I’ve been working on a way to generate Mandarin audio comprehensible input using LLMs/TTS models. The idea is that there aren’t many great CI options between 500 words and ~3k-5k words - OP himself mentions that when he started watching Scissor Seven 刺客伍六七 he barely understood anything, which is pretty hard to “push through” without some hardcore willpower. My project https://plusonechinese.com makes Mandarin audio stories that are 85% comprehensible at any level from 400 words all the way to 8k or more words and then auto-imports the audio snippets into SRS flashcards, which makes a CI workflow like this a lot easier to engage with at a lower level and without advanced willpower. Still working on making the content _truly_ interesting, but would love some feedback!

Just to offer you another perspective: In all of the (B2C) companies in which I have worked as a PM (three, across two different industries), I have been closely involved in shaping and controlling the evolution of the product.

My normal month has always consisted of creating anywhere from four to twelve unique proposals (PRDs) for new features for the company to build, with the number created varying based on the size of the company and the amount of process needed as a result. But regardless, it has always looked like: "Our users currently have X problem, and we should build Y to solve it. Y looks like this...". Yes this comes with some coordination and communication challenges, but those are always in service of figuring out what to build next.

I understand not all PM roles are like this, but I do disagree that PM never gets better than "sorting/filtering/decision maturity theatre". I am curious: if you were simply ordering an existing backlog, who was defining the new items to add to the backlog?

Earlier this week I got a warning about supply chain issues delaying deliveries while building a ford transit van in the online build & price feature

Not sure what global chip availability is, but this event made me feel like the supply chain chip chaos is alive and well

not that this is real "data" or anything, but I have lived in San Francisco (in various parts of the Mission and Soma) since 2018. My vehicles have been stolen 4 times, and recovered all four times. In all cases there was some kind of police report, but no person was ever charged with a crime. In 3 cases they actually had the person in custody and let them go (one time he ran away while being questioned in the stolen vehicle, so maybe that doesn't count, but they told me they couldn't have prosecuted anyway because he was just "in the vehicle, that doens't prove he stole it)

Separately, I have personally witnessed 3 instances of store shoplifting (like filling a whole bag with anything that you can and just walking out) where the police weren't even called because it was useless. They were just yelled at until they casually walked away.

I have watched a car window get smashed and a bag stolen from the car on Haight St. near Ameoba at 1PM on a Saturday. Everyone just watched, no one called the cops from what I could see (I didn't).

Last night I watched two guys use an angle grinder to cut a bike lock and steal a bike on Mission st. I didn't call the police.

I'm probably part of the problem now since I don't even call the cops anymore like I used to, but its hard to feel like you can do anything about it when I have seen them just let people go time and time again.

Did you try this recently with GPT-4? Here's my attempt to replicate your test:

---Prompt---

I show you a set of four magisterial unicorns in a cute pink stable, each of them has a letter painted on one side of their body and a shape painted on the other. Right now you see all four unicorns, two of which are showing the shape side of their body and two of which show the letter side. You see that the unicorns read: 'A', 'S', Triangle, Square

Which unicorns do you need to spin around to test the validity of the following statement:

"if a unicorn shows a consonant on one side of its body, then the opposite side of its body is a triangle?"

---GPT-4---

To test the validity of the statement "if a unicorn shows a consonant on one side of its body, then the opposite side of its body is a triangle," you need to look for potential counterexamples to the rule. The two cases to check are:

* Unicorns with a consonant showing.

* Unicorns with a triangle showing.

You do not need to check unicorns with a vowel (like 'A') or a different shape (like Square), as they don't violate the statement.

So, you need to spin the following unicorns:

The unicorn with the letter 'S' (to check if it has a triangle on the other side) The unicorn with the Triangle shape (to check if it has a consonant on the other side)

---

So it got the answer wrong, but in a way that I think a lot of humans would (you flip the square unicorn, not the triangle). Still seriously impressive in my book and definitely not falling into incoherence

How does this idea merge with the idea of "unbundling"? For example, I think of Craigslist, and its famous "unbundling" story[1]. For Aibnb, fiverr, Getaround, etc., the individual niches were enough to build a big product off of, and there was no need (or maybe no ability?) to grow that product back into the "everything for everyone" of Craiglist. What is different about SAP/JIRA/Salesforce? Why can't "not bad" niche-focused subsets of these products become big and profitable while still focusing on their initial niche/area?

[1] https://www.cbinsights.com/research/craigslist-unbundling/

Two economists are walking down the street and happen upon a $20 bill lying on the sidewalk. The first economist says, "Look at that $20 bill." The second says, "That can't really be a $20 bill lying there, because if it were, someone would have picked it up already." So they walk on, leaving the $20 bill undisturbed.

Just want to provide a counter-story to your anecdote. Many of my colleagues are excited to go back to the office to increase collaboration and work satisfaction. Personally, wfh is really difficult for me and I can't wait to get back in the office. Working from home feels like grinding away at an endless list of tasks and waiting for the day to be over. Working at an office feels like playing a sport with a team - every day is significantly more enjoyable and satisfying.

Granted: I am a young, single person working in design/product and my job requires a ton of collaboration. I don't really consider myself an extrovert FWIW

Not OP, but I am interested in this community and I am also getting an error when I try to sign up:

Access Denied foundersnetwork.com is using a security service for protection against online attacks. An action has triggered the service and blocked your request. Please try again in a few minutes. If the issue persist, please contact the site owner for further assistance.

I have a personal website but am not comfortable sharing here - is there a way to message you?