But because the economics don't translate the way VCs claim. When you replace a $50,000 employee with AI, you don't capture $50,000 in software revenue. You capture $5,000 if you're lucky.
"""
Good thing we solved lipid disorders with Olean, Betamax gave us all superior home video, and you can monetize your HN comments with NFTs or else I wouldn't have any money to post!
Maybe I should change the title indeed. Intention was to point to the fact that from the perspective of a startup, even if you replace it fully, you are not capturing 100x the previous market.
The year is 2012, rails is the hottest thing and mongodb is the inifinity guntlet scaling monster. BackboneJs and underscore were said to replace all jquery and we deployed things with just one command to heroku. The good ol days
You know... I was a big Capistrano fan back when I worked on Rails webapps and honestly, thinking back to it, it holds up pretty well. I can't really remember a time when I felt like it actually burned me, and that's not something I can say about very many tools. We had Jenkins automatically pushing dev builds all day long (automatically) and had a Jenkins button to deploy to prod. We did have to use the rollback functionality of cap a few times and it always went smoothly.
I learned to do deployments using custom shell scripts, and then someone introduced me to Capistrano. And i didn't get it at all. Why would i use this tool when i could just write a shell script? What does it save me? My main conclusion was that some people just don't like shell scripts as much as me.
Hah, I agree. I think the thing that made me like Capistrano was basically that someone else had written what was basically shell scripts for me and had worked out a bunch of details so that I didn’t have to on my own (eg rollbacks).
Whether cap or custom scripts, the thing I loved about that is… it’s simple enough that you can fully understand what’s happening. Not just at an abstract level but very concretely.
Duplication of keys like name and address in {"name": "Test", "address": "1234 Main St."} - it will have name and address in each record, whereas in SQL it isn't in the data, although it's in the queries.
This was the same perspective i had as an engineer working on a finance problem for accountants and sales. So i would build a nice reporting tool that had modern ui. The people liked it at first but they always regress to spreadsheet. They just ask me to export the data in spreadsheet.
"Leetcode questions are the worst form of interview questions, except for all the others" - Churchill
I too used to hate leetcode questions, but after conducting 100s of interviews as an interviewer, leetcode style questions are the best way to test intelligence and coding skills ive found in general.
However, leetcoding interviews must be paired with system design and topgrading in order to have balanced interviews. Leetcoding alone is incomplete but i find it essential.
Bias, disposition, tendency, whatever terminology you would like to give, is baked into rationale and induction. You cannot have the latter without the former. This is even more true with LLM. Without bias, LLM is inaccurate and will become useless.
This is the true 500 pound gorilla in the room. Somehow, as we all become ever more dependent on AI, we also need to become ever more aware of the biases built into that AI. Without constant vigilance, we risk becoming obedient stereotypes ourselves, oblivious to hidden biases and agendas in the smart tools that we come to blindly trust.
And modernity sometimes ignores some obvious considerations.
I am surprised that a long article attempting to find links between various manifestations of modernity to allergy, it fails to mention vaccine.
The diet, exposure to pollutants, sun, level of cleanliness etc certainly has evolved over time. But none of these have changed as dramatically, become prevalent and directly alters our immune system like vaccine does.
I don't know of any study that's been able to explain how vaccines may increase the risk of allergy. However, if we are to speculate about possible causes for allergy, it should be nearly at the top. And the omission of even a single mention is quite striking.
The fact that we cannot even talk about something so obvious should pause some of us to wonder, what is it that is causing this kind of blind spot in modernity. And that should also lead us to ask who is trying to hide this for what reason.
> The fact that we cannot even talk about something so obvious should pause some of us to wonder, what is it that is causing this kind of blind spot in modernity.
It is not at all obvious to me. For one, the timing of the introduction of vaccines does not match the timing of the allergies becoming widespread. Or are you proposing that a specific vaccine could be the link?
I've never heard of this before your comment. It certainly seems like a novel theory. Before trying to ask who is trying to hide this, why don't you propose or fund a study?
It definitely can't be the immune system altering drugs we give to children in modern society, the likes of which have exploded in usage since the early 1990s. Nope, certainly not even worth mentioning here.
But are you sure you aren't just trying to be contrarian and fishing for data?
Vaccinations were as early as the late 1800's and the first widespread programmes for yellow fever in 1930's.
Most of the allergy/asthma incidence begins to spike in the 1970's.
Excessive use of disinfectants, and excessive use of antibiotics is a much more likely mechanism.
Antibiotic resistance was precipitated by excessive antibiotic prescription in both humans and the animals humans eat.
coincidentally the increase in antibiotic resistance rising from the 1970's onwards is very much aligned with allergy anaphylatic reactions rise from 1970's onwards.
No, Lawrence Berkeley National Lab, which is Up the Hill from Berkeley. They're run by the same underlying organization but are distinct (yet overlapping in many ways). LBL evolved out of the UC Berkeley Rad Lab, run by Earnest O. Lawrence (same name as the current lab). They do non-classified research.
There is also Lawrence Livermore National Lab, which is nearby, but in Livermore. They do classified research in addition to non-classified. I suppose it's one of the two places they simulate nuclear weapons... errr, run large scale multi-physics combustion codes for stockpile stewardship.
> I suppose it's one of the two places they simulate nuclear weapons... errr, run large scale multi-physics combustion codes for stockpile stewardship.
Back in the '90s my friend (jokingly) lamented that they wouldn't let him try to play Everquest on their computer.
Japan has a unique system with lots of small police booths (Koban system) where citizens return lost objects. While we found a large number of our wallets listed on official lost and found websites, we were not able to retrieve these wallets as it would have required proof of ownership and identification documents. As a result, we were not able to find out the unique email addresses or the original locations of the wallets and we could not ascertain (among other things) to which experimental condition the wallet belonged. As we were primarily interested in the difference between experimental conditions, we had to exclude Japan from the paper. By contrast, we do not have evidence of similar systematic behavior from other countries.
I had a vision of the experimenters being prosecuted for wasting police/people's time by causing problems on purpose.
Edit: the above was a joke, but had a grain of truth. The authors addressed this specific question in a reply to readers' letters:
We did not exclude Japan from our study because their reporting rate was too low (if anything, Japan would have scored relatively high among countries in civic honesty). Since our focus was on investigating the difference in reporting rates between wallets with and without money, our decision rule was to exclude countries where individual wallets could not be identified. Japan has a unique system with lots of small police booths (Koban system) where citizens return lost objects. While we found a large number of our wallets listed on official lost and found websites, we were not able to retrieve these wallets as it would have required proof of ownership and identification documents. As a result, we were not able to find out the unique email addresses or the original locations of the wallets and we could not ascertain (among other things) to which experimental condition the wallet belonged. As we were primarily interested in the difference between experimental conditions, we had to exclude Japan from the paper. By contrast, we do not have evidence of similar systematic behavior from other countries.
Actually one of our research assistants was temporarily detained (in Kenya, I believe) for suspicious activity. As you can imagine, having a bunch of "lost" wallets on hand required some explaining.
I have lost things, or have forgotten and left minor things at various places (e.g., mom & pop food courts, shops) in Taiwan. They were insignificant in the scheme of things, and in some cases the return was more expensive than the forgotten item. Yet, they found ways to get them back to me. (A BIC four color pen?!)
This is diametrically opposite what I experienced on the mainland.
But because the economics don't translate the way VCs claim. When you replace a $50,000 employee with AI, you don't capture $50,000 in software revenue. You capture $5,000 if you're lucky. """
So you are saying, AI does replace labour.