I thought the Singularity had already happened when the Monkeys used tools to kill the other Monkeys and threw the bone into the sky to become a Space Station.
Having been in this game about 10 years longer I can understand how he feels. I distinctly remember when I realized that C compilers for the ARM produced better assembly than I could code by hand. Bitter sweet but the code being written became larger and more complex because of it.
Modern coding has become more complex than I would have ever thought possible. The number of technologies an individual would have to master to actually be a expert "full stack" coder is ludicrous. It is virtually impossible for an individual to prototype a complex Web based app by themselves. I think AI will lower that barrier.
In return we will get a lot more software - probably of dubious quality in many cases - as people with "ideas" but little knowledge start making apps. Not a totally bad thing but no utopia either. I also think it will likely reduce the amount of open source software. Content producers are already hoarding info to prevent AI bots from scraping it. I see no reason to believe this will not extend to code as more programmers find themselves in a situation more akin to musicians than engineers.
All of the hype surrounding AI will subside when a SaaS company eventually deploys a moltbot version of their software and the company is driven out of existence due to the chaos that ensues.
I wish I knew:). I kind of think Palantir is particularly at risk here. Image a company with siloed data behind APIs and access to other external data APIs. Using something like Claude I could tie all the separate sources into an easily digestible dashboard without any help from Palantir.
you'll just get replication rings in addition to citation rings.
People who cheat in their papers will have no issues cheating in their replication studies too. All this does, is give them a new tool to attack papers they don't like by faking a failed replication.
It was somewhat of a joke;) He completed Kennedy's term and and was only only elected once. He refused to run for a 3rd. Given the immense ego's of these guys one might assume it was because he was unlikely to win. Peeing on people is unlikely to win you a lot of friends.
The application of Bayesian probabilistic reasoning in general (as described in this video) is not the same thing as "Bayesian statistics" specifically, which usually to modeling and posterior inference using both a likelihood model and a prior model. It's a very different approach to statistical inference both in theory and in practice. This creator himself is either ignorant of this distinction or is trying to mislead his viewers in order to dunk on the FDA. It's obvious from the video comments that many people have indeed been misled as to what Bayesian statistics is and what the implications of its might be in the context of clinical trials.
Indeed, even more broadly online "Bayesian" seems to have taken on the form of "I know Bayes' Rule and think about base rates" as opposed to "Do you prefer jags or stan for MCMC?"
Software engineering has been on a quest to make everyone a programer for at least as long as I have been the business (>40 yrs). As that has progressed the quantity of software produced has increased and the quality decreased (with both good and bad consequences). I would expect that to accelerate with similar consequences.
1. NonProgrammers will be able to create superficially working programs for some unmet need. Good if that need does not have a large enough market to attract investment. Bad if that superficially working program exposes them to security risks or bugs that can lose them money.
2. We see a lot more BS software - Social media, Scam and Advertising. This will be due to lowering the barriers of entry since expensive competent people will not have to be involved. Expensive competent people generally have better things to do with their time.
3. New programming fad - Spec-Driven Development (SDD). IMO that will be a mixed bag. Imagine creating a spec, feeding it to Claude code one day, making a small tweak (or possibly no tweak) some n days later and getting an entirely different code in return. We will spend the next 10 years developing new strategies for just about every thing we do without really doing it much better.
4. Running our code will become more expensive and software ownership will decrease. My friend handles an extremely large corporate account for MS. Before Office 365 they did 5 million a year with MS. A couple of years after adopting Office 365 $25 million a year. AI providers will definitely do more rent seeking by convincing people that they need their service to compete, locking them in, increasing the rent.
Sounds to me like they intend to control the oil production infrastructure which is land/territory within Venezuela - but what do I know.
Isn't the entire Polymarket concept rife with ways to abuse the system? If I have insider knowledge I get shills to create a market for that knowledge - then make an extreme bet at the last moment. Seems sort of like betting the 49ers will not win the Super Bowl because you know that Purdy's kneecaps are about to be busted. Or large options trades the day before the Senate votes on Healthcare bills.
If you want a gambling site, you need to ban insider knowledge. If you want to generate accurate predictions, you want to encourage insider knowledge. But even then, the problem you mention can occur when an insider extreme bet happens at the last minute, because although you end up with an accurate prediction it isn't very useful in the few minutes before it becomes a fact. I don't know if there is a solution.
Time-weight predictions so that they're "worth" more the further in advance they are, converging to "worthless" as they approach the due date? Perhaps there is a way of making this result emerge "organically" from the rules of the system, rather than explicitly encoding it.
Depends on your goals. If you are the platform then there is nothing to solve: you’re running an illegal gambling website and currently getting away with it. If you are an inside trader you’re also doing well.
It’s not great for the gambling addicts but helping people better themselves doesn’t seem to be a theme in federal policy at the moment
Gambling sites probably do have it in their user agreements.
Further, "insider trading" in prediction markets is probably fundamentally illegal under existing commodities fraud laws in the US (I am not a lawyer,) but there's probably nobody actively policing it, and probably no precedent in how to prosecute the cases.
I think it hinges on whether "any part of Venezuela" includes intangible "parts" like being able to tell them who to sell oil to, or whether it only refers to land/territory. The second paragraph implies that control over land is the point of the bet, but it doesn't explicitly say so. Control over the oil industry doesn't require control over land.
Naive view is it's suppose to create public interest measures with real valued results.
Unfortunately, it's pretty easy to see something, eventually, like "X won't be seen in public after December 31st, 2026" essentially creating an assassination market.
reply