HN user

hermannj314

3,725 karma

Software developer and statistician.

Residence: Iowa. Luxembourg. Pennsylvania (currently).

The opinions I express are my own, but I accept the possibility I might be a Manchurian candidate sent by the algorithm.

Posts3
Comments703
View on HN

AI will also make us faster as exploitation, genocide, and weapons manufacturing.

Our technology has long since surpassed our capacity to understand its impact or to morally restrain its use - we've gone from chivalry to "oops, that was a school, oh well"

We will solve problems faster though, but we aren't showing signs that AI is making us better humans, and that is the problem.

I think it is a wonderful time to start a business. Owner-run businesses tend to have great customer service. As consumers, we have some choice to give our attention to businesses where the people doing the work are the ones that own the company, if they aren't, then give your money to someone else. Stop supporting businesses that pay their taxes to Ireland.

This isn't a perfect solution and I know there are counter-examples, but I have been much more satisfied supporting small, local or owner-run shops.

Midjourney Medical 1 month ago

You don't need to build a perfect med spa, you only need to control the patents of the key pieces of a med spa.

This is nothing more thab a prophylactic patent grab to stifle competition and progress in this space for the next decade and/or hoover up patient data behind a paywall where they will gladly lease the weights and bias cure to you own disease back to you in the form of a subscription.

I don't think a person saying Eastern European are observed doing something more than expected is inherently racist. It is a claim he either does or doesn't have evidence for.

If he made the claim with insufficient evidence or made the claim in contradiction of the evidence, then it becomes racist, but I don't think making the observation and doing the calculation is the racist part. It is a simple chi-squared goodness-of-fit test.

From what I could understand from the 6-page memorandum, (my paraphrase) "the law allows us to be nice and convenient, but doesn't require us to be nice and convenient, so we decided to make things hard and cruel going forward"

The current administration is sending a pretty clear message to immigrants.

Back in my day, when people were too lazy to put in effort they would want you to google for them, read results, interpret the data, and tell them what you found out. Lmgtfy.

Well now people are still lazy, but at least they talked to their llm, they just want you to read the result, interpret the data and tell them what you found out.

We might make better software, but we aren't making better humans.

"After my speech for the troops about how we are losing in Iran, my speech to children with cancer about how we've gutted research, sure I can then give a speech to people entering the job market about how AI is ruining the job market"

Perfect, that's exactly the message of despair we want to send! (How I imagine picking these speakers goes at every college campus)

The topic of AI surveillance pricing is almost exclusively covered by left-leaning media sources according to ground news and party affiliation largely predicted the vote with no 100% of voting Democratic legislators for and 88% of voting republicans against it.

It feels like a partisan issue.

But I guess I don't read the news.

From NASA

> Under Artemis, NASA will send astronauts on increasingly difficult missions to explore more of the Moon for *scientific discovery*, economic benefits, and to build on our foundation for the first crewed missions to Mars.

Yeah fuck me for thinking this was about that that.

A diverse and inclusive crew, a publically-funded mission, an emphasis on science and discovery, and government investment in a long-term strategy, not a quick politcal win.

This current administration has made sure these things never happen again, Artemis is very much the swan song of an America that has died. I am not interested in watching our corpse twitch and calling it life.

I have friends in the US that want the US government destroyed, there are people in the southern US that think the south won the civil war. Who cares?

Every government in all of human history has had its detractors and supporters, more detractors probably exist in expatriated communities, their existence does not really prove anything.

You might have to hold my hand on how you got there.

My actual argument was the ride-sharing addressed systemic racism that yellow cab companies did not, that is a good thing.

What net good would fentanyl do to society? If it exists, then yes, sell it. Clean, reliable fentanyl might be a good thing, I don't know.

One freedom denied to Americans is that we can not provide comfort to our enemies - this is punishable by death according our constitution, so we tend to err on unwavering support for our military always.

Many Americans may be absolutely against this horrible, barbaric, idiotic action in the Middle East, but they might wisely not want to talk about it.

So let me say "Thank you to all American troops for your service, God bless America. Our military is the only reason we have peace and freedom." - this is my official public opinion as an American and I would never have at least two witnesses catch me saying anything different.

Your question implies a sense of moral outrage about gambling so I don't if it was asked in good faith or not.

Kalshi was on my radar because of Tim Walz dropping out the Minnesota governor race and an article about who would replace him being wagered on prediction markets, coincidently this was during a boring part of January after a snow storm and so weather was on my mind.

You can wager on aggregate precipitation, high and low temp, etc. The price is a binary contract between two parties with the market taker paying a small percent fee, so the conclusion you make that it "seems like a roundabout way to lose money" was not the conclusion I drew then nor believe that conclusion to be true even now, being a market maker using a trivial model from one-day forecast data was break-even for me.

I'm not going to lie, it started as a fun thing to do on a boring and cold Saturday night after a snow storm, I was looking at weather underground map of stations around central park praying it would drop a few degrees and I'd make 4:1 on my bet. I learned a lot about weather stations in the next few weeks and it was cool looking the historical data from the Central Park weather station (I think the longest running in the US) and see how it added features and new reporting values over its long 100+ year life. It was a fun winter side quest.

I don't think this needs regulated if the people involved are responsible and having fun. No chickens died for sure, which is probably why these articles focus on the more serious bets where people are dying (and not weather).

Data is released every minute, every five minutes, every hour, and then 6 hour high/low, and then mid-day preliminary reports, so there is no last-minute since any one of those reports could contain data (with various validation/rounding caveats) that could eliminate a market of temperatures, so there isn't really a last minute.

My conclusion was to focus on forecast and attempt to predict the temperature better than forecast vs. market implied probability rather than attempt to respond very quickly to published information. I learned (with my skills) the latter was a losing proposition, but the former isn't impossible (although also possibly beyond my skills it seems).

I got into weather betting markets earlier this year since I figured those can't possibly be rigged, it is automated weather station data yet some groups in the market know the truth a few minutes before everyone else based on the way the markets move.

BUT, I stopped on the day that the PHL airport preliminary report said the low of the day was 17 and then later than day the low was raised to 18. The way the market was behaving, insiders knew the low would be retracted because normally the markets clear out a tranche of bets that are no longer possible and that wasn't happening that day.

So I don't do that. The whole game seems to be based on a group of insiders that know when and what temperature reports will say seconds or minutes before the general public and they have the capacity to play with validation on the back-end (I suspect).

I built a few models to predict weather 6+ hours out using blended model forecast data, but that didn't do better than break-even.

I don't know my point. It is the wild west, caveat emptor, you need thick skin and ridiculous attention to detail to beat the game, and even then the deck is probably stacked against you.