967 karma · joined July 10, 2023
My experience with bootstrapped companies is that they are just as subject to pressure to grow as startups, and probably have a higher success rate.
Atlassian, Github, Mailchimp and Shopify are all companies that initially started as bootstrapped companies and, in my mind, stand out as some of the most sustainably successful startups of recent history.
>Do not be so stupid to think that >A place should only belong to those who are born there
Does not imply unchecked migration since the entire social fabric of the United States is a based upon people moving here from somewhere else. Currently around 14% of US citizens are foreign born. Claiming they don't belong here would be considered an extreme view even among fairly right-wing Americans. Implying that only people born here belong here is, in fact, a absolutely 0 immigration view, that would essentially destroy the United States since our population growth is almost entirely from immigration.
Likewise:
>It is not okay to say >Build a wall to keep them out
US immigration has not resorted to wall building as a form of immigration control for all but the most recent years. Again, nobody I know who is opposed to wall building supports "unchecked migration".
The statement there only reads to me as a call to have compassion for those caught up in what will be a perpetually escalating migration crisis (and will likely soon make migrants of those who are currently protecting their borders).
> Pretending the development of a programming language is inexorably linked to unchecked immigration is disingenuous.
Where am I (or anyone else) making that claim? I'm claiming cross-cultural/international cooperation and mutual aid are unquestionably tied to the development of not just a programming language, but all open source software. Anyone who has worked in this space at all can likely list someone they have worked with on nearly every continent.
The fact that you view calls for compassion and empathy as calls for "unchecked migration" is a bit concerning.
It's also worth pointing out that Python (and it's ecosystem) is developed by an international team of people, most of whom volunteer their time out of good will and sense of shared purpose across diverse cultures.
The "both sides" logic seems to miss the point that world without mutual aid and international, cross-cultural cooperation would be one that would not have Python and it's ecosystem. It's not political for the Python developers to support such a view, such a view is foundational for the existence of Python.
Work spilling over into extra days does happen during crunch times, but is much rarer. Though I likewise fear that 4 day work weeks would simply mean you end up working 8-8 everyday to get the same amount of stuff done.
I do find that my philosophy and critical theory books are the ones that more heavily come from university presses. In general the humanities tend to have stronger presence in the university presses (largely because books, not papers, are the publication format of choice for those areas).
Haven't we seen this type of thinking to be proven a fallacy enough in the last decade to at least be skeptical of solving that last 1% until we actually see it?
It's worth pointing out that, historically speaking [0], we're not an in era of particularly high interest rates yet, just not absurdly low ones.
While I do believe AI (imho shorthand for LLMs and Diffusion models) has a lot of potential, it's main function has been to reignite irrational exuberance in the tech industry.
I work in this is space and I'm very curious how long the mania can hold out, and even more curious what the consequences will be for the people seriously working on AI when it dies down.
Layoff didn't stop because a correction had finally concluded, but because investors snorted another line of hype and were happy to throw their money around again and pretend everything is fine.
The come down from this, I suspect, will be pretty rough (but hey, if I'm wrong it's great for me).
The premise that there is something optimal about the state we evolved in has no basis in what we understand about evolutionary processes. Those are just the conditions that we happened to not die in, not the conditions that are in any way "best" or "optimal" for us. It's like sleeping in late and missing a flight that ends up crashing and assuming that the optimal strategy is to always sleep in late.
Further, in an ever changing environment, which is what make evolution so important, "history" in a genetic sense is completely uninteresting. Things that cause one species to not die in one era can easily be the source of it's extinction in a future one. Megafauna probably had many benefits, pre-human existence, from being very large. But being large and reproducing slowly is what ultimately let to their extinction once the environment changed.
To be clear on my point: "Nature" is not a rational agent with a plan. No creature evolved a certain way for a reason, they just evolved that way and happened to not die when some others that didn't evolve that way did die.
Nature doesn't even "intend" for you to pass on your genes, it's just that that's how the system happens to work such that things that pass on their genetic information still stick around.
The very premise that there is some "natural" thing that we would be better off doing, silently assumes a pretty major, border line theological, premise that there is any reason that living things should continue to a path that allowed them to survive to a point. If evolution has any "lesson" teach at all it's closer to "be flexible".
I saw a reviewer mention that VR is like "having your favorite amusement park ride at home": it's awesome, but at the same time you're not going to ride a roller coaster for 3 hours everyday.
VR can be loads of fun, but I still find it exhausting, and very focus demanding. It will, for me, always remain a nice plus, but never as essential as something I can easily pause and put down.
I'm sure the internet will be fine, and the web has already been essentially destroyed as the drive for extracting revenue from every human interaction has rendered it just an amusing advertisement for the most part.
Most of the content of the web today is already generated by "bots" even if those "bots" happen to be human beings.
I hate the idea of having to use a mouse to click on a visual GUI to navigate a file system in order to make use of its functionality.
It's less the case today, even among developers, but it wasn't that long ago that I remember that any serious technical user of a computer took it as a point of pride to touch the mouse as little as possible. They're also still correct in that thinking. The command line is a very powerful UI with lots of benefits and while the mouse makes navigating the OS easier it's still much more limited than command line usage.
Touch screen interfaces are another example of an easier UI that ultimately feels even more limited. But people still plug their iPad pros in to magic keyboard folios frequently.
Having worked with these tools everyday for awhile now the "AI will change UX" is such a better take than "AI will conquer the world!". AI does fundamentally open up new work flows and user experiences, many of which do over a lot of potential improvements over their predecessors.
At the same time I doubt we'll see a world where we don't end up using the command line for the majority of serious technical work.
Except that doing so requires more energy than you got out of the process which created the CO2 in the first place.
The only way to mitigate climate change is to keep CO2, in the form of Hydrocarbons, in the ground in the first place.
If we have any hope at averting the worst case scenario we would be start globally marking proven reserves as "never to be extracted". As it stands we are on course to release all of the CO2 stored in currently leaded fossil fuel reserves. We can slow that process down a bit by regulation, but it will matter on the scales we're working with.
However before it breaks down methane is a wildly more potent greenhouse gas. It still takes decades to break down, during which time it's going to be multiple magnitudes more impactful that the CO2 it will eventually become either way.
I've seen this phrasing frequently in regards to carbon sequestration and it always seems to wildly understate the real problem with removing CO2 from the atmosphere.
CO2 isn't some unnecessary byproduct of creating energy, compared to say soot in wood burning. The CO2 created is an essential part of the carbon cycle by which all living things store and release solar energy. CO2 + H2O + (Solar) Energy => Hydrocarbons + O2. This is how energy is stored on our planet from the energy in a high calorie soda to petroleum (okay technically sugar is not a hydrocarbon, but the idea still holds).
When we extract energy from foods, burning organic material, or fossil fuels we do so by reversing that equation: Hydrocarbons + 02 => CO2 + H2O + Energy
To "scrub" carbon necessarily requires more energy than we got out of the process in the first place. Photosynthesis, for example, is only 4% efficient. Which means it takes about 25x as much solar energy to build the log you burn then the heat and light you experience burning the log.
"Scrubbing" CO2 fundamentally requires tremendous amounts of energy, and more than we got from the energy source in the first place. There are natural processes, like rock weathering, than can do this without energy inputs, but those are hard to replicate and scale.
China serves as this counter factual and it's facing an even larger housing crisis. Massive building of houses lead to many people buying up property for investment reasons and now there is a pretty clear oversupply while plenty of people still struggle to find housing.
I don't really understand why HN is obsessed with building unlimited housing as though this will fix the problem. For years people clamored that this was the solution to all of SFs problems, but as SF starts to shrink it already looks deserted, I can't image what happen if the HN fantasy of building unlimited massive apartment housing had materialized.
I wouldn't use China as an example of solving the housing problem by simply building. There's currently a large oversupply problem [0] and, not entirely unlike the US (but more extreme), excess supply of housing has lead to people simply owning more than one property as an investment strategy.
0. https://www.wsj.com/articles/chinas-smaller-cities-struggle-...
> you just heard through the grapevine that it's happening
Is that "grapevine" public? If not, it is by definition material non-public information.
In general in finance if you think you've found some clever exploit in the system that you're surprised no one else is taking advantage of, it's a good idea to double and triple check that it is in fact a legal exploit. Especially if you stand to gain millions off of something that seems easy to do.
You are only able to trade share's during specific windows of time typically 1 week after earnings are released.
Additionally virtually every company I've known has explicit policies stating that you cannot buy/sell any derivatives related to the company stock (which is a shame since buying put options is a legitimate way to insure your compensation).
Further more, even these rules are only this lax for non-executive or other high level employees. If you're higher up in the company you have much more access to non-public material information. The solution to this is usually to set up a 10b5-1 that automatically liquidates shares based on a schedule approved by the board.
In regards to the "previously" question. I wouldn't worry about legitimate trades, but if you are trading based on insider information and looking to gain a lot of money, then trading would, by definition, be "insider trading"
Sure it's not great at differentiating between SotA techniques, but it's very useful for sanity checks like this one.
Even for SotA models, it's still useful to verify that you can get greater than 98% accuracy on MNIST, before exploring larger, more complex bench marks.
It certainly shouldn't be the only benchmark but it's a great place to start iterating on ideas.
Most people don't realize that Logistic regression can get ~90% accuracy on MNIST.
As a big fan of starting with simple models first and adding complexity later, I've frequently been told that "logistic regression won't work!" for problems where it can in fact perform excellent.
When faced with this resistance to logistic regression I'll often ask what they think the baseline performance of it would be on MNIST. The guesses I hear most often are 20-30%.
People, even machine learning people, often don't realize the rapidly diminishing returns you get for adding a lot of complexity in your models. Sometimes it's worth it, but it's always good to start simple first.
It's also been my experience that if you don't get good performance with a simple model, you are very unlikely to get great performance from a more complex one.
McKinsey is in the business of reputation laundering. They start the process by building up reputation. They hire from prestigious schools, have prestigious customers, make themselves a household name for the people that matter.
Then when you need someone to make an unpopular decision for you, or take the blame for a project failure, they swoop in and take you money in exchange for taking the reputation hit. It's fine, they know how to grow more reputation when necessary, and everyone at your org or in the media if the situation calls for it, points to them and says "what a bad/stupid/down right evil decision McKinsey made!" while remain the responsible person who tried their best by hiring what you and your peers thought was the best.
If people are making more money relative to inflation and the workforce participation rate has been steadily rising, then why is credit card debt at record highs and personal savings at near record lows? That should imply that people are buying tons more stuff than they used to, because if you make more money, don't save the money and you still need more credit where does the money go?
There were and increasingly are some major concerns about the economy. For starters we're still seeing very low unemployment with rapidly decreasing personal saving rate (near record lows) [0], increased tapping into 401ks early[1], and record credit debt [2].
More people are working and yet are less able to pay for things. Even if you completely ignore inflation, something seems very concerning about this.
The fact that the market is divorced from reality doesn't mean that our economic fundamentals are strong.
That doesn't even touch on the looming crisis with commercial real estate impacting the already unstable small banks which is just starting to unravel now [3], or the major threats to the insurance industry that are starting to appear with increased effects of climate change [4].
What's going on right now economically is strange, and it's a far better to question "wait, why aren't we in a recession?" than it is to mock those who claimed we would be (though I do tend to avoid 100% claims).
[0] https://www.finder.com/personal-saving-rate#:~:text=The%20pe....
[1] https://finance.yahoo.com/news/workers-are-still-raiding-the...
[2] https://www.cnn.com/2023/08/08/economy/us-household-credit-c...
[3] https://www.wsj.com/real-estate/commercial-real-estate-regio...
[4] https://www.wsj.com/finance/insurance-catastrophe-reinsuranc...
What benefit would sound, thoughtful leadership provide in a world where the fundamentals have no connection with real world success? For years if you had made a profitable, viable and sustainable company investors would have scoffed at you. You would have been crushed by some investor hype fueled, unprofitable, impossible to sustain company chasing trends.
Reality keeps threatening to rear it's ugly head, but then we find a way to keep it at bay for another round of delusion. We all know something is not quite right but as long as there's money on the table we all keep playing the game. I mean we all know by now that crypto is a scam and COIN has more than doubled it's opening price at the start of the year.
From your questions it looks like you are only interesting in the R part. RAG implies the retrieval step is then used to augment a user prompt.
To answer 1, a good heuristic would be "can a human reasonably familiar with the terminology answer questions about the meaning?" If a human would need extra info to make sense of your data then so would an LLM.
This is where RAG typically comes in. For example if you had documentation about ClassName and FunctionName, a retrieval model might be able to find the most likely candidates based on a file containing full definitions of these classes and function, then pass that info into the LLM appended to your query.
For 2: It depends if the fire house is the query or the data. If you have queries coming in very quickly, then you might be able to if your firehose doesn't have too much volume since you can batch requests and get responses fairly quickly.
If the fire hose is the data going into the vector DB then you might have some difficultly inserting and indexing the data fast enough.
Honestly, as someone who has been very concerned about climate for a while now, I'm surprised how much visible disruption of the climate system we've directly been able to observe. Earlier in my life I thought that, though dire, this was certainly a problem that would impact future generations but not so much ours. It turns out I might have been quite wrong on that front.