News for hackers. Stuff that matters.

thehoneybadger

Pretty disappointed with Hacker News allowing bigotry onto its platform. I've been here for over a decade. Count me out. Enjoy mixing politics onto your platform. The mind virus has truly spread far and wide.
Re:How about trailing commas in SQL? on Tuesday February 11, 2025 @12:48PM (#43015767)
I nominate this post for worst idea of the century.
Re:Re-Evaluating GPT-4's Bar Exam Performance on Tuesday June 04, 2024 @01:35AM (#40571013)
Yes, that is an accurate comparison.
Re:Re-Evaluating GPT-4's Bar Exam Performance on Sunday June 02, 2024 @04:39PM (#40557071)
It is difficult to comment without sounding obnoxious, but having taken the bar exam, I found the exam simple. Surprisingly simple. I think it was the single most over hyped experience of my life. I was fed all this insecurity and walked into the convention center expecting to participate in the biggest intellectual challenge in my life. Instead, it was endless multiple choice questions and a couple contrived scenarios for essays.

It may also be surprising to some to understand that legal writing is prized for its degree of formalism. It aims to remove all connotation from a message so as to minimize misunderstanding, much like clean code.

It may also be surprising, but the goal when writing a legal brief or judicial opinion is not to try to sound smart. The goal is to be clear, objective, and thereby, persuasive. Using big words for the sake of using big words, using rare words, using weasel words like "kind of" or "most of the time" or "many people are saying", writing poetically, being overly obtuse and abstract, these are things that get your law school application rejected, your brief ridiculed, and your bar exam failed.

The simpler your communication, the more formulaic, the better. The more your argument is structured, akin to a computer program, the better.

As compared to some other domain, such as fiction, good legal writing much easier for an attention model to simulate. The best exam answers are the ones that are the most formulaic and that use the smallest lexicon and that use words correctly.

I only want to add this comment because I want to inform how non-lawyers perceive the bar exam. Getting an attention model to pass the bar exam is a low bar. It is not some great technical feat. A programmer can practically write a semantic disambiguation algorithm for legal writing from scratch with moderate effort.

It will be a good accomplishment, but it will only be a stepping stone. I am still waiting for AI to tackle messages that have greater nuance and that are truly free form. LLMs are still not there yet.

Re:What UI density means and how to design for it on Tuesday May 21, 2024 @03:21PM (#40432661)
The Visual Display of Quantitative Information is one of my favorite coffee table books. Props for a Tufte reference!
Having written and sold machine learning software for over 15 years, you are definitely over-reacting. There is about a 10 year pattern. Every 10 years AI gets drummed up as the next big thing. It is all marketing hype. It always fails.

This happened in the 70s, 80s, 90s, 00s, 10s, and now the 20s. Without fail. It is a hyped up trend.

Only be concerned when someone is presenting a real breakthrough in the science (not the commercial aspect). A real breakthrough in the science will not have any immediate economic impact.

Convolutional neural networks are absolutely not revolutionary over the prior state of the art. These are incremental gains in model accuracy at the cost of massive data structures. There is no real leap up here in the ability for a machine to reason.

ChatGPT is just snake oil. Calm down. It will come and go.

The well deserved bad rap for MS is waning. They made some changes in the last few years that seem to be a step in the right direction. I will give them the benefit of the doubt. For now.
Re:Ask HN: Learning modern web design and CSS on Monday September 25, 2017 @08:14PM (#15335211)
Do the same thing you might do when browsing open source code on github. Go to websites of well known companies or by people that provide web design services, view source, look at a feature that doesn't make sense at first, and then search on Google or Mozilla Developer Network or Stackoverflow for more details, and then write a practice page in a basic text editor.

As far as design, you cannot go wrong with bootstrap. However, to really learn design, do not study web design. Study things like Bahaus, billboard design, marketing and messaging and PR, how to compose for simplicity, SEO optimization techniques and their influence on design, etc. Seriously, the people that learn how to make websites and want to tell you how to make websites absolutely and hilariously SUCK at design.

Generally, when a visitor visits a website, which I call an interaction, then they must agree to its terms of service. This forms a contractual relationship. Contracts are private, and outside of things like not promising to do something illegal, it is up to each party to decide on its terms. In this case the website generally decides on the terms, and then the visitor opts into them through use of the website. Much of the issue lies within contract formation. When a lawsuit happens later, one thing that is frequently questioned is whether the contract is valid. A common angle is to claim the the contract is invalid because it was not correctly formed. Bad contract formation. For example, in a different setting, you didn't sign your name on the dotted line when buying a house or something like that, so there was never a contractual agreement, so there was never an actual sale, etc.

However, there is case law (law made by judges who hear cases and issue opinions) that says that sometimes contracts can be implicitly formed. For example, if you as a website visitor are given proper notice of a website's terms of use, and then you continue to the use website, you have implicitly agreed to the terms of the use. Even if you didn't sign anything, or check any box somewhere saying you agree. No explicit action has to take place.

Except, that is not exactly worldwide statutory law (laws passed by government and written down in the books with codes like Law #1234.56). While the issue of formation is mostly settled, there is still some room for creative legal maneuvering. Aka lawyering the shit of things. Aka screwing things up because someone with deep pockets is paying you to win using any angle you can get.

This cookies notice and agreement probably falls right into this category. And while it is generally settled law that the contract is formed even without this agreement, some schmuck somewhere still thinks there is wiggle room, but it is merely case law and not exactly authoritative, especially not in the international setting.

When in doubt, lawyers adhere to CYA. Cover your ass. Use the narrowest, most conservative, safest interpretation of the law. In this case, there is this tiny bit of doubt, so CYA. Just in case.

I personally believe you can make a good argument that contract is implicitly formed merely from continued use, and the notion of requiring express consent is outdated. The law is catching up to how things are done online, the trend is rather obvious, and anyone whining about it is probably just some established cash cow business that somehow wants to manipulate the market to further extend its antiquated business practices and is willing to spend millions on dollars on go screw yourself legal teams.

So, yes, you could theoretically get in trouble. But you are not likely to, and anyone suggesting otherwise probably has an ulterior motive.

Re:Ask HN: Is Bloch's Effective Java Still Current? on Sunday June 12, 2016 @10:16PM (#11891372)
I think it is still current, and is a decent read. Like, you should know these things, but not necessarily adhere to them. I spent a decade believing in OOP. Lately, my thoughts have changed. Unless you really need to track non-global state, or need a namespace in a language that doesn't have namespaces, I don't think you need to go down the OOP route as far as Mr. Bloch advocates. Sometimes a simple function just gets the job done, clearly and effectively.
Re:Ask HN: What are your favorite scholarly papers? Why? on Saturday November 29, 2014 @09:45PM (#8675802)
Wow you beat me to it! This paper affected my college major, my career in my twenties, and my current career. Such an important paper, although, I must admit, it is not easy to decipher at first.
The question is a bit open ended, so here are a few thoughts. Programming involves learning by doing. Just start on something. Perfection is the enemy of the good. Your first real attempt won't be perfect. In fact it will be downright terrible. Set your expectations accordingly and do not let this prevent you from getting started. It will also never be finished. Do not let the scope of the project prevent you from getting started. In fact you might even be bad at estimating the amount of time it will take. Ignore those fears and confusion and just start. Channel your weekend warrior instinct. Sit down on a Friday and just start typing stuff.

Start with an idea, a problem. It really helps when the problem is personal. What do you hate? What annoys you? The text editor you use? Some little feature of JS that the makers of the language seemed to have forgotten about (how could they!)? Some website that just looks wrong to you? The todo list that doesn't work the way you want? Just think of something that doesn't quite work the way you want, a void, and then try to fill it. Just start and get in over your head. Do some spying on how others solved it (preferably only if you really get stuck, or once you made substantial progress). Maybe you can solve it better? Gotta dog food it. I find dogfooding a solution to a personal annoyance is what keeps you progressing towards a finish line.

If your asking about, say, the idea of website design, and the typical systems that make up sites (the software stack), just start reading about website design. Make a dumb website. Make a blog. Make a family photo album. Basically, start, and then figure out what you don't know, and then figure out what you need to know to proceed. It is an iterative learning process of realizing all the things you do not know and calmly tackling them. For some, this is the best part, the adding of another tool to the toolkit, the unearthing of some common pattern among the tools.

Re:Why programming methodologies are pointless on Tuesday November 25, 2014 @04:44PM (#8659901)
I got to admit, for a second I thought this was another meme that I did not know about. I need to trust in myself more.
Ugh, speaking of UIs, this UI has a pest problem. Dedesign this site before promoting it.
Re:Ask HN: Ways to increase mental sharpness or intelligence? on Saturday August 16, 2014 @02:01PM (#8186945)
Never skip over the definition of a word with which you are unfamiliar. For example, acuity.
Re:Ask HN: Who isn't in the software industry/not a hacker? on Friday August 08, 2014 @05:03PM (#8155019)
Evil corporate lawyer.... love code, algorithms, new tech, and insightful commentary. HATE advertising.
Your asking the question and stating its puzzling is why (I mean this in a civil way, my rant is directed at the general populace).

Technical people typically use a technical approach to solving a problem, which is not always the best approach. Technical people generally make poor historians. Decisions are a-posteriori, through personal experience and personal failure. Knowledge must be internalized so it becomes intuitive. But this very step is an act of specialization. Expecting the non-specialized market to understand, on faith, this masking of technical complexity, is strange. See, e.g., object-oriented design patterns, geek-speak, a hacker-news comments echo chamber, etc.

There is relatively little value placed on documentation vs performance/correctness. Products are rarely user-centric. There are too many black box APIs. Dumb questions are frowned upon. Gross assumptions are made.

"New" products are quite often not actually new. They are merely evolutionary, not revolutionary. They merely add to the market, or slightly expand it, instead of defining a new one. Just another person internalizing some jargon and slapping a price on it and then proclaiming PEBKAC as if he/she said something insightful. This is not innovation.

The value to creators is to learn a new skill set, merrily hacking away at trees, and chopping down the right tree using brute force luck, a rather naive traveling salesman. For some reason the hard problems are considered easy and the easy problems are considered hard. The difficulty of achieving simplicity is under-appreciated. You solve some technical problem, and stop there, thinking you have won. But this is not the end, this is the beginning. The problem never was minimizing big-o or memory efficiency. The problem was, and is, messaging; the communication of the solution.

Investors recognize market saturation and look at evolutionary innovation as low hanging fruit. "Killer" products are harder to create and higher risk. Demand is more fixed in evolutionary space but demand can be uncovered when entering revolutionary space. The problem is not Wile's proof of Fermat, it is explaining the value to everyone. Explanation is the messaging channel. Properly encapsulating the complexity of a problem is the difficulty. Properly recognizing the problem is the difficulty. Implementing it is the easy part. Much of nerd culture (including myself) has this backwards, through no fault of their own (INTPs are introverted, formal analysis is prized).