A large part of the original deep blue win was smoke and mirrors - humans researched how it should play and hardcoded it in, humans researched the specific openings it should play based on what they decided would work best and hardcoded them in, humans tweaked it / changed its programming mid-match to fix bugs and get better results, the moves it decided on its own were based on pure brute force analysis of positions, no "thought" involved, humans waged psychological warfare against Kasparov to get him to play worse, just massive smoke and mirrors. And yet the public knew none of the details and just gobbled up the hype / result as "computer smart, ai beats human".
Now we've got a new system of smoke and mirrors and hype and people that don't understand what's actually happening under the hood, don't actually understand the technical details of what's involved, just fully buying into the smoke and mirrors put up by people making billions of dollars by convincing people of the hype.
Hint: Make sure the people giving you the efficiency improvement numbers don't have a vested interest in giving you good numbers. If so, you can not trust the numbers.
Reminds me of my last job where the team that pushed React Native into the codebase were the ones providing the metrics for "how well" React Native was going. Ain't no chance they'd ever provide bad numbers.
>> As a relavant example, try using FreeType in your C/C++ project, make sure your solution compiles on Linxu, and Mac, and Windows (and ideally other platforms)
You know, as a (prickly) analogy, whatever your take on covid was, half the population vehemently disagreed with your take. No matter which side was more "correct", either way, a huge percentage of the population can be, and often is, completely deluded on even fairly understandable topics.
When it comes to something as complex as AI, what are the odds that a random person is going to have any sort of good/informed take on it? Especially someone like this, who's a non-technical angel investor? Their entire job is hyping things up to raise money / get paydays. They actually list on their resume various "viral articles/tweets" that they made that got attention / raised money. Could this guy remotely explain, technically, how an LLM works under the hood? I highly doubt it. His credentials are not building AI, not technical knowledge, but hyping up companies that use AI.
Well, at least he gives 1-5 year time frames for all his grand claims, so when they don't actually happen he'll be quickly proven wrong. But of course, it's the internet, and nothing will ever come from somebody making grand claims and then being completely proven wrong, there will be no follow up, no self reflection, no retraction, no long-term credibility hit, just on to hyping up the next thing after getting his payday.
It feels oddly freeing to be seeing headlines like this every other day on HN and not caring in the slightest. The titles are just amalgamations of random words to me, like 'Claude Super Zen Deep 4.0' or 'Grok Hyper 6.2 Mega'. They come and go. A month from now it'll be new headlines with new numbers and new words. And I still won't care. Not in the rat race, just using whatever chatgpt gives me for free. Just coding how I've always coded.
But hey, they raised rates numerous times to pay for all those nonexistent new features. I cancelled my membership a week or two ago after yet another rate hike.
Ok, let me guess, without looking at the article .... is it a "pilot" that's rolled out to a small number of people, for a limited period of time, and its success is judged by surveying those people on whether they were happy to get free money? I bet it was.
Yes, this is the problem. They tout this new latest and greatest extension that fixes and simplifies a lot, yet you go look up the extension on vulkan.gpuinfo.org and see ... currently 0.3% of all devices support it. Which means you can't in any way use it. So you wait 5 years, and now maybe 20% of devices support it. Then you wait another 5 years, and maybe 75% of devices support it. And maybe you can get away with limiting your code to running on 75% of devices. Or, you wait another 5 years to get into the 90s.
You know, I've been following a rule where if I open any article and there's meme pictures in it, I instantly close it and don't bother. I feel like this has been a pretty solid rule of thumb for weeding out stuff I shouldn't waste my time on.
I like how they make grand claims and when you finally drill down into the data you find it's based on something like ..
"We took the opinion of this random bank about how efficient it is to allocate capital in the country, which translates to how corrupt it is"
or
"We had this (ironically, corrupt) NGO make a report where they asked some unspecified number of "credentialed" economics professors at an unspecified number of unspecified universities what they think about corruption in their country."
Yes, those are actual examples of the data/methodology they provide. That's the type of data backing up their ratings.
"Eight more months of Bitcoin. It's usage continues to dramatically expand. The amount of transactions is increasing exponentially. Soon, fiat currencies will collapse, all replaced by Bitcoin transactions. If you haven't converted your assets over to Bitcoin you're going to be left behind and lose it all. I can't even understand people that don't see the obvious technical superiority of Bitcoin, such people are going to go through rough times."
" They can write code better than you or I can, and if you don’t believe me, wait six months."
It's ALWAYS wait 6 months, or wait for the next generation. Or "oh that model that we told you to use is old now, use this new one instead. Oh that doesn't work? Well that's old now, use this one". Always. 6 months ago it was wait 6 months. 12 months ago it was wait 6 months. 18 months ago it was wait 6 months. Now it's wait 6 months. 6 months from now it'll be wait 6 months.
In addition to never providing examples, the other common theme is when you dive into the author's history almost 100% of the time they just happen to work for a company that provides AI solutions. They're never just a random developer that found great use for AI, they're always someone who works somewhere that benefits from promoting AI.
In this author's case, they currently work for a company that .. wait for it .. less than 2 weeks ago launched some "AI image generation built for teams" product. (Also, oddly, the author lists himself as the 'Technical Director' at the company, working there for 5-6 years, but the company's Team page doesn't list him as an employee).
Ah yes, to have AI write code for you, you simply just need to, let's see ..
"Document the requirements, specifications, constraints, and architecture of your project in detail. Document your coding standards, best practices, and design patterns. Use flowcharts, UML diagrams, and other visual aids to communicate complex structures and workflows. Write pseudocode for complex algorithms and logic to guide the AI. Develop efficient debug systems for the AI to use. Build a system that collects logs from all nodes in a distributed system and provides abstracted information. Use a system that allows you to mark how thoroughly each function has been reviewed. Write property based high level specification tests yourself. Use strict linting and formatting rules to ensure code quality and consistency. Utilize path specific coding agent prompts. Provide as much high level information as practical, such as coding standards, best practices, design patterns, and specific requirements for the project. Identify and mark functions that have a high security risk, such as authentication, authorization, and data handling. Make sure that the AI is instructed to change the review state of these functions as soon as it changes a single character in the function. Developers must make sure that the status of these functions is always correct. Explore different solutions to a problem with experiments and prototypes with minimal specifications. Break down complex tasks into smaller, manageable tasks for the AI. You have to check each component or module for its adherence to the specifications and requirements."
And just like that, easy peasy, nothing to it.
As a supreme irony, the story currently on the front page directly under this one ('You are here'), makes the claim "The cost of turning written business logic into code has dropped to zero. Or, at best, near-zero." in the very first sentence.
Utter rubbish from an extremely biased source. Every time they say something like "didn't you notice X" or "your portfolio must look like Y" the answer is nope, you're completely wrong. Every time they talk about some major "crash" you can just go look at it and see that it recovered within 48 hours and looked identical to dozens of other events through recent history. The outright calls for violence + intentionally destroying our own economy to "stick it to the man" at the end surely make me believe this is some rational analysis.
You know, I was expecting what the post would say and was prepared to dunk on it and just tell them to stop using ai then, but the builder/thinker division they presented got me thinking. How ai/vibe coding fulfills the builder, not the thinker, made me realize that I'm basically 100% thinker, 0% builder, and that's why I don't really care at all about ai for coding.
I'll spend years working on a from scratch OS kernel or a vulkan graphics engine or whatever other ridiculous project, which never sees the light of day, because I just enjoy the thinking / hard work. Solving hard problems is my entertainment and my hobby. It's cool to eventually see results in those projects, but that's not really the point. The point is to solve hard problems. I've spent decades on personal projects that nobody else will ever see.
So I guess that explains why I see all the ai coding stuff and pretty much just ignore it. I'll use ai now as an advanced form of google, and also as a last ditch effort to get some direction on bugs I truly can't figure out, but otherwise I just completely ignore it. But I guess there's other people, the builders, where ai is a miraculous thing and they're going to crazy lengths to adopt it in every workflow and have it do as much as possible. Those 'builder' types of people are just completely different from me.
Ok cool, so in a single year when this hasn't happened, we know never to listen to any grand claims he makes ever again.