HNHacker News
TopNewBestAskShowJobs

loveparade

2,076 karma · joined December 30, 2022

Meet Me At The Love Parade
submissionscomments
loveparade··on Absence of superconductivity in LK-99 at ambient conditions
Sure, but you really think peer review would solve any of that? It would only make things move 1000x slower and be worse.

The current "free-market" process of replication attempts is going to resolve the issue both quicker and with more statistical significance than peer-review ever could.

Perhaps this isn't what you're implying, but the peer-review word in your comment jumped out to me, as if that changes anything.

loveparade··on Most promoted and blocked domains among Kagi Search users
I've been a paying Kagi user for about a year and didn't know about this feature! That's amazing, I can finally block all these spam sites completely.
loveparade··on Polars: Company Formation Announcement
I am worried about VC money specifically, not money in general. There are other ways to fund OSS projects that don't lead to perverse incentives. Plenty of projects make money through sponsors, partnerships, or additional products without VC incentives.
loveparade··on Polars: Company Formation Announcement
Oh no, that doesn't bode well. Congrats for them for getting funding, but personally I've learned to stay away form VC-funded OSS projects. Sooner or later, they need to focus investor returns, which inevitable results in locking more and more features and maintenance behind "products" - Of course, they never start out with that intention and promise that they'll always focus on OSS. Never happens.
loveparade··on Kraken Technologies: How we organise our large Python monolith
Why does it matter? Just because I don't like Google doesn't mean they suck at writing Python code. The same goes for anything crypto.
loveparade··on Ask HN: Anyone else notice that HN isn’t full of JavaScript frameworks lately?
AI "frameworks" like langchain have overtaken the space JS frameworks took previously.
loveparade··on LazyVim
A bit off-topic, but I wonder how much Copilot/GPT cuts into the market share of editors like vim and emacs. I used emacs for the past few years, but I recently switched (back) to vscode because the Copilot integration is really good. There are a lot GPT integraton packages for other editors, but they don't come close to the deeper integration that vscode has. And of course, vscode will probably get priority treatment from Microsoft.

At this point, this is (somewhat unfortunately, I really like vim and emacs) a big selling point for vscode.

loveparade··on Hacking LangChain for fun and profit
That's a bit cynical. It does many other thing as well. For example, it exposes simple re-usable modules, also known as "functions" in Python. Sometimes it also uses Python "lists" by wrapping them into... classes.
loveparade··on Ask HN: Build Own PC in 2023?
Are you saying that I don't need to smolder and put thermal paste manually like I did 20 years ago!? Time to build a PC.
loveparade··on Show HN: An Ethereum-authenticated community on Web3, science and technology
In this case, there is nothing more anonymous about the Ethereum approach than a simple username/password. But there are some benefits in terms of data privacy. Most websites, especially social media, don't let you get away with just an anonymous username/password these days if you want to use their full functionality. They may want all kinds of other information from you, e.g. email address (3rd party), your phone number, ID verifcation, linked Google account with certain permissions to access private data, and so on. This data will be sent to their servers (as well as ad providers / Google) and linked with your identity. You have little control over this and can't remove it from their servers either.

With the Ethereum approach, you can control this data about yourself. For example, with zero-knowledge proofs you could prove to a service that "I am over 18 and an US resident with a valid phone number" without sending any actual information about yourself to that party. Twitter shouldn't need your phone number. It just needs to know that you have one.

Logging in with your Google account everywhere is convenient because it contains all kinds of facts about you. But you have no control over what's done with that data, either by Google or the service you're logging into. With "Log in wit Ethereum" you could get the same conveniences, but you still own your data and have full control over what's being done with it or who can access which parts of it.

loveparade··on How to Succeed: Lessons from Sam Altman
I think your definition of success is just different. If you define success as moving upwards and having a very comfortable life, I agree with you.

But if you define success as becoming super-rich with your startup and investing, that's mostly what the poster above you says.

loveparade··on How to Succeed: Lessons from Sam Altman
Be at the right place at the right time when you buy your ticket. But psst, don't tell anyone.
loveparade··on Wash trading gone wrong – how a $100M/day crypto exchange bankrupted itself
I don't think bankrupt is the right word here. It's the opposite. They didn't bankrupt themselves. Their scam was exposed at some point so they stopped doing it. They didn't lose any money. For what we know, the creators may have been quite successful in extracting money from its users, including Bertyar88.

So, the more interesting question to me is how much money BaseFEX actually made from this scam before they turned it off.

loveparade··on AI and the Automation of Work
I don't think this is a good analogy. A better analogy is the multiple previous AI hype cycles. They played out similarly. They brought some truly new capabilities that are used in production, just like LLMs will become a widely used tool, but failed to live up to the AGI hype even though everyone was convinced we're just a step away back in the 80s already and in a few years machine will do all our work.

The only reason people don't see this is because they weren't around back then. It's not going to be different this time.

loveparade··on Behind the Pretty Frames: Diablo IV
The hardware being limited is not an excuse. It's of course much more work to polish high-fidelity graphics than it is to polish something on an old console where you only have a few pixels to work with. Zelda looks and runs abysmally compared to any modern non-Switch game. People should stop giving Nintendo passes for that.
loveparade··on The idea maze for AI startups (2015)
I don't think it is. The distribution of an AI model that was trained on such a huge amount of movie reviews is very close to the human distribution.

At least that's true around the mean. If your application needs to handle long-tail cases, an LLM won't easily give you that. But depending on the application, that may not be necessary. So yeah, sometimes this is a bad idea, but for many applications it may be just fine.

loveparade··on Hexyl: A command-line hex viewer with colorized output
Even if it's part of the training data and the LLM is just a better search engine, how would I have figured out what the code does without an LLM? I certainly can't paste this into Google.

I mostly agree with the stochastic parrot interpretation, but that doesn't undermine the usefulness or impressiveness. Even if it's just a highly compressed search index, that level of compression is amazing.

loveparade··on “But the SEC let us go public” and other flawed arguments in Coinbase's defense
Nothing wrong with expressing an opinion, it's just sad to consistently see these highly biased and political opinions being upvoted on a website that's supposedly about rational thinking, not political incentives. I would expect upvoted articles here to strike a fair balance in presenting both sides of the argument, not just one.
loveparade··on “But the SEC let us go public” and other flawed arguments in Coinbase's defense
It's not about what's wrong, it's about the selection bias of what's included in such an article. You can write a technically "factually correct" article by only including statements that heavily favor one side of the argument, which is what's done here.

Also, "facts" in topics that touch politics and law are complicated. This is not like science where measurement can be unambiguous. Most facts in topics such as these like these are not scientific facts, but highly politicized opinions that are ambiguous and open to interpretation. Much of the legal code is (on purpose) open to interpretation, including what's a security and what's not.

When trying to understand a topic like this the IMO best thing you can do is seek our content from people who have a somewhat neutral stance. Don't listen to a16z who want to pump all their crypto tokens. Don't listen to Molly White whose whole career depends on bashing anything crypto-related.

loveparade··on “But the SEC let us go public” and other flawed arguments in Coinbase's defense
I don't understand how anyone can take this obviously super-biased articles from someone who has made it their career to hate anything crypto-related seriously. It's full of personal jabs and bias like "my guess is that it’s going to involve a lot of PR stunts and headline-grabbing"

It's obvious that there is zero incentive here to do any real reporting or research. It's just clickbaity headlines and statements appealing to people who hate crypto. Of course it's hidden under a veil of neutrality with a biased selection of links that nobody actually clicks on scattered throughout the article.

loveparade··on htmx
It's no different from anything else. The promise if every framework is to abstract away some low-level details. Whether that means writing "higher-level JS" like jQuery did many years ago already, and hundreds of other frameworks do today, or inventing some new DSL like htmx does that translated to JS internally. No difference.
loveparade··on htmx
htmx is another tool just like that, another goalpost in the frontend webdev world. It's just happens that the most recent hype is "don't write JS", but fundamnetally it's no different. Just another framework/library that has gotten some traction. In a few years we'll be back full cycle and it will be in vogue to write JS again, probably because htmx and associated projects have grown to complex.
loveparade··on The unreasonable effectiveness of character-level language models (2015)
They are. Nobody in the AI research community expected that scaling up models several orders of magnitude would make them that good. Even the original selling point of the Transformer paper was "can be much bigger and faster than RNNs, even if not as good" - Of course everyone was wrong and it's all about the data and scale.

Btw, tokens are nothing new in modern models. They've been used since 2015/2016 [0]

[0] https://arxiv.org/abs/1508.07909

loveparade··on Bard now open to use
So far it's pretty bad for coding. I tried 10 questions I gave to GPT-4, which got all of them correct. Bard gets 0 out of 10 correct. Significantly worse than GPT-3.5 even.
loveparade··on Please don’t upload my code on GitHub
No, the difference is that OpenAI has a huge competitive advantage due to direct partnership with Github, which is owned by Microsoft. In fact, it's even worse. With OpenAI making money from GPT, Github has even less incentive to make data easily available to others because that would allow for competition to come in. I wouldn't be surprised if Github starts locking down their APIs in the near future to prevent competitors from getting data for their models.

Nobody is arguing against uploading code. It's about Github/Microsoft specifically.

loveparade··on Please don’t upload my code on GitHub
A key difference is that a company is making a proprietary paid product out of the learnings from your code. This has nothing to do with open source.

If the data could only used by other open source projects, e.g. open source AI models, I don't think anyone would complain.

You could argue "well, but anyone can use the code on Github" and while that's technically true, it's obvious that with both Github and OpenAI being owned by Microsoft, OpenAI gets a huge competitive advantage due to internal partnerships.

loveparade··on How are you using ChatGPT internally at your company?
The use case where it's really good is boilerplate code that you may not remember. For example, I haven't written React code for years now. I know what good React code looks like when I see it because I've worked with it before, but I wouldn't be able to implement something from scratch without googling or copying stuff from Github. I just don't remember the libraries and best practices off the top of my head.

When I ask ChatGPT to do it for me it gives me an excellent starting point. Sure, there will be bugs, but because I know what I want I can spot and fix them immediately. It's much faster to adjust ChatGPT's code than it is to Google around for starting points.

Another example are shell scripts. I only touch bash once every few months and I keep forgetting the syntax for certain operations. Asking ChatGPT to give me a starting point is much faster than googling and visiting 20 StackOverflow posts for what I want.

But I agree with you that for day-to-day work on the same codebase where you have all the context, ChatGPT usually isn't worth it.

loveparade··on Re-implementing LangChain in 100 lines of code
Am I the only one who is not convinced by the value proposition of langchain? 99% of it are interface definitions and implementations for external tools, most of which are super straightforward. I can write integrations for what my app needs in less than an hour myself, why bring in a heavily opinionated external framework? It kind of feels like the npm "left-pad" to me. Everyone just uses it because it seems popular, not because they need it.
loveparade··on The age of average
You know what's also the same? All the products in the same category on Amazon.
loveparade··on Smart urinals stuck on boot at a gas station in Germany
It's fine, Google killed that product a while ago. You probably got caught in the middle of the shutdown.
← PreviousPage 10 of 13Next →