3,729 karma · joined October 8, 2013
is there a risk of bias? when VCs make decisions, there is no hindsight. but when you ask humans to evaluate predictions, you necessarily are using past (training) data and the humans may recognize general successful patterns or even the examples themselves.
For example: the model outputs that it's considering an investment in a company that lets drivers pick up passengers on the way to their destination and earn some money as well. a person may think, "Duh, this is uber! invest!" Thus inflating "success" rate.
Maximizing ad profit on search traffic necessitates more paid ads, both on search results page and the web pages. They won’t say that directly of course but that’s what we’ve seen play out over last 20 years.
Vague metrics like “good user experience” get trumped by hard metrics like total $ / search.
The reason larger companies write prds, figma mocks, etc is to get alignment across more people that are detached from your individual thing
At a small company you don’t need that. Everyone is working on one main work stream
At larger companies, there’s hundreds or thousands of important projects at one time and lots of supporting teams for those. Everyone has their own shit to worry about and can’t possibly remember all the context on your projects, so you make it clear as possible to quickly bring disassociated people up to speed on why it’s important and what you need from them
It also helps disassociated leadership buy into your ideas. A picture is worth a thousand words.
That’s not inherently evil is it? It sounds pretty nice.
The evil comes in people squeezing others. Whoever make the robots obviously have a lot of power, corporations would quickly find ways to cut out the “parasitic” humans and get the robot labor, etc.
Can spam provides for up to $50k PER EMAIL in civil penalties.
If you make 1 cent or $10 per email, doesn’t matter. It’s no where close to that level of penalty. So you make damn sure you don’t ruin yourself.
Now we just need that kind on text messaging - it’s a Wild West these days
1) Turn on filtered view on iMessage
2) Actually report the abuse to carriers. iOS makes it easy, but it seems pretty ineffective because the abusers can just use another number. But if you complain to the carrier directly, then they can (hopefully) remove you entirely for that shady customer (and possibly kick them off). Here's what I do
a) Go to https://www.ipqualityscore.com/free-carrier-lookup (or whatever site you like, that's just the one I found)
b) Type in the spam number
c) Find the carrier name
d) Google the carrier, go their site, and find "Report abuse" or something similar
e) Fill out the form. Include your contact info so you actually know whether something is done or not.
At some point it was done and I got my discounted theme park tickets or whatever.
But I can easily see how the high pressure sales process works on people. They have a bag full of manipulative tricks.
The thing about people is they’re people. They’re different. It can be hard. They’re selfish. But there can be joy in our differences, and real meaning when someone actually cares. Vs a chatbot who is centered around you. It might feel good but it’s not a real relationship of give and take.
the job is to make YouTube videos that people click and watch
What gets them to watch and stick is a few things but notably wow factor, something crazy they haven’t seen before
The bar for wow factor keeps rising
Therefore you need to keep learning driving better and better results. Otherwise you are out
You need to take ownership for results to avoid delays at all costs.
Let's assume today a LLM is perfectly equivalent to a junior software engineer. You connect it to your code base, load in PRDs / designs, ask it to build it, and viola perfect code files
1) Companies are going to integrate this new technology in stages / waves. It will take time for this to really get broad adoption. Maybe you are at the forefront of working with these models
2) OK the company adopts it and fires their junior engineers. They start deploying code. And it breaks Saturday evening. Who is going to fix it? Customers are pissed. So there's lots to work out around support.
3) That problem is solved, we can perfectly trust a LLM to ship perfect code that never causes downstream issues and perfectly predicts all user edge cases.
Never underestimate the power of corporate greediness. There's generally two phases of corporate growth - expansion and extraction. Expansion is when they throw costs out the window to grow. Extraction is when growth stops, and they squeeze customers & themselves.
AI is going to cause at least a decade of expansion. It opens up so many use cases that were simply not possible before, and lots of replacement.
Companies are probably not looking at their engineers looking to cut costs. They're more likely looking at them and saying "FINALLY, we can do MORE!"
You won't be a coder - you'll be a LLM manager / wrangler. You will be the neck the company can choke if code breaks.
Remember if a company can earn 10x money off your salary, it's a good deal to keep paying you.
Maybe some day down the line, they'll look to squeeze engineers and lay some off, but that is so far off.
This is not hopium, this is human nature. There's gold in them hills.
But you sure as shit better be well versed in AI and using in your workflows - the engineers who deny it will be the ones who fall behind
His presented focus is on great experiences for the customer, which demands high operational excellence from his team. His team found a reasonable way to save time, but his reaction was that's the wrong focus. Focusing on saving time will eventually hurt the customer experience (like advertising models lead to enshittification), even if this particular one doesn't.
So they need to re-orient this - either go back to the hard way, or justify the easy way in terms of the customer. To quote, "they wanted to sandbag it, they needed to figure out how they would make up for the lost energy elsewhere."
I have other areas I’d rather ask about, and travel id ask more basic questions - where was your favorite trip in past few years? Where do you most want to go? From there can go deeper on why, what they prioritize to do, how they approach traveling, etc.
I’d add a feedback box to the AI. Click random prompt, then I can tell the AI “no travel questions, give me some questions to find what they value in life”, or whatever, then can continue to tweak from there
Atheism has always existed, it just may have been more or less visible depending on the political climate.
Even long ago when people worshiped ideas like Thagwag the rain god[0], there were probably non believers - are we so arrogant as to assume they didn’t exist?
“Hm we pray every day, sometime rain sometime not. Maybe no Thagwag?”
[0] made up
One of my favorite episodes was his first homicide case. A gas station was robbed and the clerk was killed. No video, no prints, etc. However there was a necklace that was dropped. It had some kind of pattern, so he called every. single. jeweler in the area. He called a lot, and eventually got a hit and found his man.
It's this kind of diligence and dedication (plus technology) that can crack cases open, even cold ones. Like the article here, he found 4 untested hairs that cracked it open.
As a manager I’ve had two employees tell HR that I was racist. The evidence? One I fired for performance, the other I had on a performance improvement plan. Mind you I had other minorities on my team in parallel that had no performance issues and strangely enough did not say I was a racist.
Also one time the HR guy (who also doubled as office manager) ran a large scheme where he claimed employees were expensing things, he did it on their behalf and got reimbursed. I found this out after the fact where I was asked if I ever asked him to order laptops or ran up huge Uber bills.
Example here
The Chancellor interrupted to ask whether he could cite any case that permitted common law ratification of “an adjudicated breach of the duty of loyalty.” Ross danced around the question, insisting that Defendants were not arguing that the vote changed the Court’s adjudication “with respect to conduct.” Blah blah blah, skip a bit...
“So,” asked the Chancellor, “this has never been done before?” After a bit more rhetorical contortion, Ross finally admitted that, “Yes,” this has never been done before.
Thus, with these questions, the Chancellor established that for her to rule in Defendants’ favor, her Court of Chancery, which is subordinate to the Delaware Supreme Court, would be placed in the position of establishing a completely new legal doctrine.
It's like a brilliant magic trick, if you can follow along.
“By contrast, taxation is a notable exception in effectively causing large emission breaks alone. It stands out as the only policy instrument that achieves near equal or larger effect size as a stand-alone policy across all sectors.“
Uber got its start by flouting taxi rules. By the time the backlash started uber/lyft were too popular for governments to shut down so instead most governments created an awkward different category for them, so there’s at least some regulation applying to them.
If Uber had tried to work thru the regulation they would have never gotten off the ground or have been as loved by users. Taxis in many ways suck, and that was their opening.
Speaking of which…Uber is a master of regulatory influence. When California threatened to classify their drivers as employees, they successfully passed a proposition (prop 22) with the California public that promised more benefits for drivers…but kept them as independent contractors.
It's kind of like planting trees - it can take a while to see the fruit, but then it keeps on giving with some ongoing maintenance.
In my life, my significant other, my career, and financials are those areas I heavily invested early on, and a decade on they are quite strong. I still struggle, but when I take a step back, I can see these are very good and feel grateful.
There were many times I have been tempted to leave. But I kept an eye on the big picture, what would leaving really change for me? And the answer was not much, I am still me (as noted in the article).
On the other hand, I have under invested in myself (health, hobbies, spirituality) and friends. My main struggle is working to shift my focus away from maximizing the former, and begin balancing more of the latter.
It's an ongoing journey :)
The survey seems to have missed a likely large confounding variable - cheating.
Men are more likely to cheat, so their female partners, either due to confirmed or suspicious cheating, may demand location access.
It explains why every tech product is eventually doomed to shittification and opens the doors to competition
Value creation is when you have a new thing and people are adopting. You grow by more people using. Think Google photos offering free unlimited photo storage.
As some point you saturate the market. Then to keep growing you shift to value extraction. Going back to example above, now photos/videos count toward your 15 GB Google limit.
Sundar appears to be an excellent extractor, hence why Google has never done better financially but completely missed the boat on AI. AI is now the new wave of value creation.
Kind of like you may hear a language, and recognize what it is, but no idea what they are saying.
The Navajo language was not written at the time, and it’s very different than others and subtle differences in tone can make all the difference. Their written notes of what they heard were probably generally inaccurate which made breaking the code impossible (vs enigma which was exactly copied, and the output is German, they just then needed to break it)
I like the follow up questions feature but how is it different than chat gpt - just providing links as well?
The better it does as an investment class, the worse it will become. It's a self perpetuating spiral.
I don't see the market solving this. I only see some regulation helping here - ie any tax filer can at most own 1 SFH (this includes married couples). So no second, third, tenth homes. Corporations and foreign entities also cannot own SFH - bye bye black rock & China.
Obviously some edge cases don't work (ie people who want to put their home in a trust), but generally speaking, people should not be able to own homes as investments.
Their prices are generally too unstable to put anything on "ship & forget". There are some exceptions, like the items you subscribe to.