HNHacker News
TopNewBestAskShowJobs

mlinsey

5,217 karma · joined April 18, 2008

www.twitter.com/mlinsey

Engineering @ Homejoy (YC S10)

hnchat.com:8pJMixG6jypNc7WE7714

submissionscomments
mlinsey··on The YC Winter 2022 Batch
doing just a quick search in the public directory, I found 24 out of 394 companies are in crypto. That's a lot lower fraction than you'd expect given how much mindshare, hype, and attention crypto has now vs. other sectors.

I am a crypto-skeptic, and consciously chose not to work on it myself, but I still wouldn't want any investor, especially one as early stage as YC, to categorically rule out a sector based on the impact of its current technology, the state of the technology will be different when these companies have grown.

Eg with respect to the carbon impact of crypto specifically, the current chains are very unscalable for the exact same reason that they consume lots of carbon; while the per-user stats look bad now, any world in which these companies are used at the scale of web 2.0 platforms is necessarily a world where the carbon footprint per user is many orders of magnitude lower.

mlinsey··on Fire extinguished at Ukraine nuclear power plant, Europe's largest
Your link supports the comment you are replying to; it says that after the reactor is shutdown, additional cooling is needed for some time, and that eg the Fukushima and Three Mile Island incidents both involved damage done by an inability to cool the reactors after they were already shut down.
mlinsey··on Boeing’s 737 Max Software Outsourced to $9-an-Hour Engineers (2019)
"under the hood" can go arbitrarily deep; nobody thinks everyone needs a deep understanding of the quantum physics governing how a transistor works to build a web app.

But in my experience, people saying "under the hood" are using it with respect to the level of abstraction they are working with. There's a major inflection point in effectiveness when you go "one level deeper", or understand how the tools you are directly using are implemented, even though there are certainly many layers of abstractions below those tools. It helps a ton with debugging and performance work, and you start to really understand the limitations of your toolchain.

mlinsey··on Ask HN: What did Marc Andreessen mean by this?
Open platforms, where anyone can build anything without asking a central authority (with their own financial incentives) for permission, allow for much more innovation. As an easy example, take a look at the App Store and Apple's 30% tax, banning apps which are too "similar" to others, and compare it to the web, where anyone can try most any new idea and potentially disrupt incumbents. Should be noted that before pmarca was a VC, he was an entrepreneur at Netscape, so he knows the benefits here and has lived through opening and closing of the web.

web3, to steelman it, has the potential to get us back to that permisionless world, and in fact go even further, by allowing data to travel with users. At 10K feet I like to think of web3 as "what if instead of one central Facebook with everyone's data, there was a global shared DB of everyone's social content, people logged in with the keys to their specific rows of the DB" In theory, this could bring about a golden new era of social apps - instead of counting on FB to provide the right API endpoints that you needed for your app, and then not banning or copying you, you could build true first-class clients to people's existing social data, with the exact same DB access that the original platform builders had. This would allow for all sorts of new social apps that users would love but aren't themselves viral or funded enough to re-create the whole social graph from scratch. Incidentally, it would also be very disruptive to the FANG giants and put much more control directly in users' hands.

I think it's important to take this upside case seriously, even though I am actually a big web3 skeptic. I think, in particular:

a) the scalability concerns are inherent to the tech (and PoS etc have their own issues)

b) introducing so much financial speculation on top of the space is discouraging people from doing the hard work of building real useful apps when you have these liquid tokens and can cash out just for hyping up a project, before you have product-market fit.

c) a global shared DB means the data model will advance and iterate at the speed of eg, the HTML standard, not at the speed of a startup; this means ambitious apps will just build their proprietary side data models to give users the features that aren't approved by the core spec, and if they capture enough of the mindshare, those features then become expected by all users, and apps that don't have them become second class - the third party has now taken an open data model and turned it into something proprietary. (this is basically the "embrace, extend, extinguish" strategy we saw Microsoft use against open standards in Web 1.0) You are already seeing signs of this with eg OpenSea adding other metadata on top of NFTs

d) most fundamentally - I think we're kidding ourselves how much the average user likes having a centralized authority to appeal to. Central authorities can ban the person harassing you from the site, reverse a credit card transaction that you authorized under false pretenses, stop supposedly-academic organizations from getting your permission for your data and then giving it to Russian propagandists, even censor ideas you don't like. Crypto early adopters are individualist libertarian types who think of all these features as unnecessary or even outright harmful, but I think a lot of ordinary people would rather be on central platforms with these things.

mlinsey··on Why skyscrapers are so short
A moderate spread of six story buildings might be optimal for energy usage of just the buildings, but total energy usage/environmental impact are also shaped by other human factors for where people want and need to live and work...eg without tall buildings in Manhattan, you would get a lot of extra energy expended in more people commuting back and forth from further distances. And the a high density of people makes public transit even more viable which brings further environmental and health benefits.

That said, I do agree that there are only a few areas in the country where the main barrier to an ideal density is allowing more skyscrapers instead of mid-rises. Manhattan north of the financial district and south of midtown is one such area, parts of SF are another. But many, many more cities would see more benefits of moving from low density single-family homes and detached lots to mid-rises.

mlinsey··on NFT's aren't the answer to the problems of digital art
Ahh, in that case, the problem this solves is the database owner has admin power, and could take away your graphics card.

This is similar to how I was always worried that Bank of America would arbitrarily change the balance in my account for some reason, but bitcoin solves this problem. Now I just have to worry about losing my password or getting hacked and losing all of my money, and that's just much less realistic than a major bank abusing their centralized admin powers to change my balance.

mlinsey··on New MacBook Pro has first ‘DIY-friendly’ battery replacement design since 2012
I disagree. It's true that the management style he had during the 1980's definitely would have had a stronger and faster backlash in the 2020s - he would have been forced out of the company even sooner than he really was. But this style also didn't work all that well in the '80s and led directly to him being ousted and Apple entering a slow death spiral they had to be rescued from a decade later. His second act really was different and he was much more mature as a leader; certainly he still worked people hard and was unusually blunt compared to some other current FANG CEOs, but you didn't see the sort of petulant and intensely personal abuse that you would sometimes see from him in his first act.

I think if he was CEO of Apple in the current environment, certainly there would be less hero worship outside of Apple, and a lot more people would be writing pieces critical of his leadership style and effect on the world, but there wouldn't be anything serious enough to force him from the company given how well things would be going. In other words, he'd be much closer to Musk than Kalanick both in terms of the severity of the criticism and his ability to weather it.

mlinsey··on Ask HN: Best way to host a website for 500 years?
Other comments have given ways to physically archive the webpage. Continually hosting it is a much trickier endeavor. Beyond just keeping the servers up, technologies will shift such that eventually html webpages, servers that talk using tcp/ip, datacenters that connect via fiber cables, etc will all be deprecated.

That said, if we have a very liberal definition of the word "website" to include any successor technologies where a device can be used to request a document, given an identifier, that looks recognizably like your webpage, this is doable. What you really need is an institution that you can trust to keep existing and to keep the necessary upkeep of your website as part of its mission.

The main institutions I can think of that have lasted for 500 years unbroken are churches and elite universities. If you were able to convince the Pope to decree that the church should keep hosting your webpage in perpetuity, that would likely work, but persuading him of that sounds very challenging. That said, universities are used to accepting gifts with sometimes eccentric strings attached. The gift will probably need to be large; but I imagine a $1B donation to Harvard under a condition that they continue to host and update the page as needed would likely work. Getting that sort of money is quite hard, but tbh probably easier than coming with a way of guaranteeing that your direct descendents keep the webpage up.

mlinsey··on [dead]
Podcasts feels like one of those apps Steve wouldn't have used, so it wouldn't have seen as much polish. I suspect this is why Keynote is fantastic but Pages and Numbers were unusable (at least, the last time I attempted to use them regularly, about 10 years ago)
mlinsey··on Intuit to Acquire Mailchimp for $12B
I dunno if I agree with this. If you're early on in your career, jumping in as employee #20 can be an optimal time to learn a whole lot by working directly with all the key senior people in the company. If you're later on in your career, an employee #20 is likely jumping on board to work with people they know and trust.

From a strict compensation perspective, you are probably right, but even then, it is probably less stark than it was ten years ago, what with seed rounds and series A sizes being so massive that salaries are a bit more reasonable at this stage too.

mlinsey··on Tesla opens a center on Native American land, selling cars straight to consumers
They have said that only Tesla can best explain the best advantages of their cars, a similar justification that Apple gave when they opened up their own retail stores. And while I can now buy Apple laptops at Best Buy, I do usually pick Apple stores and prefer that experience.

I will say as a consumer, I dislike the typical car dealership user experience. Lots of back-and-forth games on pricing, getting upsold on random add-ons or financial products, having to wait around a lot while your rep checks in the back about yet more paperwork, etc. Maybe some people feel this bazaar-like experience helps them nab a good deal, but I doubt I'd be worse off paying a standard price set by the manufacturer, delivered to me by a store that's not trying to squeeze out their own margin through those add-ons.

mlinsey··on Woman sues L.A. after being struck by car on a street where tents block sidewalk
Mental health is a large part of the issue, but sometimes cause and effect can be murky.

It's important to note that even the very large visible encampments in cities like SF and LA are the tip of the homelessness iceberg, there's a much larger group of people who are living on friends/relatives couches, in their cars, or in shelters. A lot of these people have some level of employment, but housing affordability (a result of ~50 years of very restrictive land-use policies making it hard to build new housing) means that this level of employment can't afford them their own permanent home. For example, prior to 1970 it was much more common to see "boarding houses" where you'd share a room or share a bathroom, and especially unmarried low-wage workers would live there, but these are illegal to build in most cities these days.

Most of this homeless-but-not-on-the-street population is transient - it is common eg to be temporarily living in your car after a job disruption or sudden emergency expense, then to find another job and find a place to live after some time. In other words, they will do exactly what you are suggesting and lock down a new job quickly. And this group is a majority of the group of people who spend any time homeless in a given year.

But a portion of of this transiently-homeless population doesn't solve their job situation quickly, will wear out their welcome with friends, or will lose their car for a variety of reasons -- and end up falling into street homelessness. Once you are homeless and living on the street, you are at extremely high risk for developing a mental health and/or substance abuse problem even if you didn't start out with one. It also becomes much harder for you to get one of those available jobs, for a number of reasons, some as basic as not having transportation to work or not being able to present yourself well for an interview.

In short: land-use policies cause homelessness, the majority of the people who experience any amount of homelessness do get a job quickly, the street-homeless you see are the small fraction (of the frighteningly-large homeless population) that got stuck for some reason, and their situation is a feedback loop making them even more stuck.

(also: the 'we will hire you on the spot!' thing is a weird 2021 artifact of the pandemic, it taking time to ramp back up, and the super-generous unemployment benefits program that will be expiring very soon. the hiring market won't be like that next year)

mlinsey··on AI Dungeon public disclosure vulnerability report
I find either take here to be plausible, is there any good research? (I can't imagine how to ethically run an experiment, but maybe someone's figured that out)
mlinsey··on AirPods Max
"Does your solved problem bluetooth headphones RELIABLY connect to your current device (without having to unpaid with another device). That is game changing."

Man, I wish my headphones did this. I constantly have to fiddle with my bluetooth settings, have audio come out of my mac or iPhone, sometimes in inappropriate places, even after I've tried to tell the OS to play from headphones. The worst is when switching between my Mac and iPhone, usually I have to unpair one, put my headphones into pairing mode, or at the very least turn bluetooth off on one device.

My headphones are...Apple AirPods. They did this when they first launched, so I even got the pros. Now none of my pairs ever work.

I really wish Apple would figure out how to make stuff just work. I think there's opportunity for someone to figure it out! No clue what happened to this company, but the AirPods are undoubtedly the most frustrating Apple product I've ever owned

mlinsey··on Say goodbye to hold music
For booking complex, non-fungible services (like Doctor's appointments, plumbers, etc) phones are great for high-bandwidth communication for follow-up questions, triage, etc. Phones also allow a certain element of salesmanship which doesn't come through on a form, which a big reason why service providers like them.
mlinsey··on PG and Jessica
If YC were a non-profit, they definitely wouldn't have been able to raise money from LP's to scale the program, and would therefore been able to accept far fewer teams than they do now, certainly not everyone!

I agree that there's nothing wrong with wanting to make money AND help people, but I also want to add that strictly maximizing the positive impact does not mean strictly minimizing making money either, sometimes making money enables you to help a lot more people too...

mlinsey··on Facebook sues EU antitrust regulator for excessive data requests
You're generally correct, but there are exceptions. The payor still has the right to see claims if they want to, and sometimes they will carve out certain services from the insurance plan that's administered for them, usually for cost optimization reasons.

In a famous example, Wal-Mart carved out back surgery from their healthcare plan and instead negotiated directly with a few specific hospitals across the country, where all Wal-Mart employees who need back surgery then travel to for care.

More common is for big companies to contract with primary care groups to serve their employees directly. (The thinking is that incentives can be better there, since some primary care doctors stay afloat by selling their practices to large hospital chains that they then are encouraged to be aggressive about referring their patients to, raising total costs of care. I think Facebook has this sort of on-site primary care arrangement...this is particularly common in very high cost-of-living cities where it's hard to have what most would think of as a doctor's lifestyle as a primary care physician.

mlinsey··on Who’s behind Wednesday’s epic Twitter hack?
After a Twitter employee suspended Trump's account on their last day of work in 2017, I expect access to his account to do anything at all is much more highly restricted within Twitter's internal tools.
mlinsey··on Hackers take over prominent Twitter accounts in simultaneous attack
I'm guessing you work/have recently worked at a big tech company (FANG or one of the ~5 other companies of comparable size) and are seriously overestimating how common their best practices are. Unless by "passing grades as an undergraduate" you mean "bonuses and promotions at a majority of the companies that handle your data every day"
mlinsey··on Hackers take over prominent Twitter accounts in simultaneous attack
I do think it's odd that so many prominent accounts were hit but not Trump's. I remember there was an incident a couple years ago that a trust and safety employee at Twitter suspended Trump's account on their last day. It's very likely that after that incident, special guards were set in place to prevent admin tools from messing with Trump's account. This would align with speculation that this hack targeted an internal employee admin tool.
mlinsey··on Hackers take over prominent Twitter accounts in simultaneous attack
Yes. Also, we're about an hour in now, and Musk's account just sent out another tweet after the message had been posted and deleted several times. At this point, if it was just an account compromise, someone would have reset it by now
mlinsey··on New Facebook tool allows employers to suppress “unionize” in workplace chat
I agree with the other reply to you, but to address your point directly, it is probably true that a lot of the frontline workforce can't afford to leave, but the people in leadership positions and very high-level/well-respected ICs still have options (including, given given FB's pay and recent stock performance, the option of not necessarily needing to work). That's a small group but each departure would be noticed and they'd have plenty of leverage.
mlinsey··on Citing revenue declines, Airbnb cuts 25% of workforce
Anyone who loses health insurance due to losing their employment (even if they voluntarily resign) is legally allowed to buy COBRA for 18 months. I think they are paying for it, otherwise this would be a fairly meaningless statement since the employer isn't involved with COBRA at all and that's less eligibility than any employee would have.
mlinsey··on Show HN: Dashboard for tracking insider trading at S&P 500 companies
I feel like most quarters won't necessarily give easy opportunities for insider trading, because stock prices over the course of the quarter are greatly moved by macro trends that are hard to forecast. But there are definitely some short-term events that move a stock in fairly predictable ways (an acquisition offer, an imminent recall, earnings coming in way way above or below guidance, etc). With a bit of leverage it's easy to just wait and make plenty of money off of particular events rather than maintaining a continuous insider-trading portfolio. Of course this strategy makes your trades much more suspicious too...
mlinsey··on Covid-19 twice as contagious as previously thought – CDC study
The main study which backs up his guess is the report from the WHO's trip to Wuhan, where they actually did what was suggested above and mass-test random samples from the population, finding a very low incidence of positives. Personally I would find the WHO on a trip to China as not the most reliable of sources, so take it with an enormous grain of salt. But it's one data point.

Meanwhile, your link in [0] flat-out says we don't know the percent of asymptomatic cases. It is well-documented they exist. I have also seen some papers estimating the percent of cases that are asymptomatic as well under 20%, and as high as the 80-90% range, the difference between them making an enormous difference to the implied mortality rate.

A month ago, people pointed to South Korea's 0.5% fatality rate, a country which tests very aggressively, as an indication that the true fatality rate is much lower. Well, they still test just as aggressively, and now their mortality rate has risen to 1.96%. The disease started spreading in a population that skewed young there, and but now the disease has started to hit older demographics who have a much higher mortality rate. And of course, "asymptomatic" cases can also mean pre-symptomatic cases, as happened with the Diamond Princess, which had an observed 46% of positive cases be asymptomatic...only to have this rate fall to ~17% later.

mlinsey··on Why Is Facebook Not in the Cloud Business?
this is true, but a lot of resources put into a new team that was completely walled off from the rest of the company, in terms of roadmaps and eng culture, could overcome this problem.

I think someone at Amazon in the mid-2000s would not have necessarily looked at their internal stack vs. other companies and picked them to become the breakout cloud winner, but when I was there (late 2000s), the AWS team was kept quite separate, and aside from things like S3 the retail side of the company wasn't really using much in the way of AWS infrastructure yet.

mlinsey··on Italy and South Korea virus outbreaks reveal disparity in deaths and tactics
Most hospitals don't have all that much slack in bed capacity; especially at the tail end of flu season most of those beds are already filled. The hospitals may be extrapolating the growth rate of the disease and realizing how soon they will have problems. Or given the nature of the outbreak it could be that even within Lombardy there are areas with much higher concentrations of cases, like there are more cases, inc. per capita, in Wuhan than other cities in Hubei.

What's not plausible is ICU beds being taken up by people who don't have the disease (maybe miild-to-moderate flus or bad colds) and are just panicking. Emergency room lines, sure. Test shortages, absolutely. But hospitals will not put someone who's just scared and doesn't even need to be admitted at all into the ICU.

mlinsey··on What Will Happen in the 2020s
I think this is inevitable if #1 is true. It may seem politically infeasible now, but if governments get really serious about climate change, then the price of many meats but especially beef will increase significantly, and that alone will give the fake beef a cost edge.
mlinsey··on Nuclear fusion startup raises $100M to design and build a demo power plant
In an ideal world, government would invest in many more low-probability/high-impact engineering projects. But this doesn't align well with political incentives - 100M on a bet that will return 50X 10% of the time is great for a private investor, but a politician won't get 50X the votes from a successful project as an unsuccessful one, instead there's a 90% chance the project will be portrayed as a boondoggle.
mlinsey··on Duolingo becomes first VC-funded $1B Pittsburgh tech startup
There is also a monthly subscription plan. Dunno what their conversion rate is, but recently (maybe sometime this year?) they added a game-style resource mechanic where any mistake makes you lose a heart, which you gain back either by waiting a few hours or by being a premium user.
← PreviousPage 4 of 22Next →