Moreover, the tests should be randomized somehow to ensure the models don't memorize the answer.
22,047 karma · joined September 4, 2011
Moreover, the tests should be randomized somehow to ensure the models don't memorize the answer.
It was garbage before LLMs got good, and the deep thinking models leapfrog whatever it used to be in its heydey.
Emphatically yes.
Just as freely as you use the words and language of dead people.
One day you won't even think about it. Well, your descendants won't. The aversion dies with you.
Squeamishness over thinking and speaking is pretty rare. Visual creation is just a new dimension of our conversation, and it'll be as accessible and common as writing words down.
You know the old generation called the world wide web satanic? They died and we stopped worrying about it.
What about when poking fun at Hitler, such as Kung Fury and Jojo Rabbit do? Off limits?
Can one paint Bob Ross in a sci-fi setting? Is that unethical? Is it okay so long as a "real" artist does it and not an "AI artist"? Or can we seriously never depict him again because someone might get angry?
"Mr. Rogers in a blood-stained sweater"?
Come on!
And we can't animate Bruce Lee? Or put him on stickers? Or have him in new movies? (Nevermind how Quentin Tarantino did him wrong.)
When I grew up Elvis effigies were everywhere. And I mean everywhere. The people who liked this stuff are mostly dead or retired now so you don't see it as much, but it used to be a big deal. Really big deal.
So we can't do that anymore?
Abraham Lincoln hunts vampires for Christ's sake...
In case folks aren't aware, Disney tried to avoid an allergy-related wrongful death lawsuit from a restaurant at Disney Springs (large publicly available outdoor mall that doesn't even require tickets) because a couple had a Disney+ streaming trial:
Insiders have an edge and can in theory shape the overall distribution of bets such that the real world outcomes are priced in better.
I'm pretty sure "an historic" is the norm in American media, not the other way around.
"A his-toric day"
"An his-toric day"
Both seem to work correctly with American pronunciation.
Journalists and films seem to use "an historic" with an American accent, which breaks the rule but still feels correct.
For most words the indefinite article is clear. This one feels weird.
I suppose that prevents Kalshi gambling on box office duds?
I prefer to prototype in Golang since it compiles fast and makes for quick iteration, but at the end I ask the LLM to port the Golang to Rust.
Nice Rust enums for APIs are the chef's kiss.
I absolutely will not write anything in Python or scripting languages anymore. They're too brittle and don't have great devex or deployment stories. Especially when you can just as easily build in a typesafe language with good error handling that compiles down to a single static binary.
It's not just AI, it's everything. Romantic rivals, celebrities, sports teams, fandoms, choice of IDE or programming language.
It's a social evolutionary strategy to neg and gossip about people and things you don't like or are jealous of. It's low energy, has little downside, and can reduce the fitness of rivals, especially when done in group settings.
The algorithm has co-opted this part of our primate brain and turned it into ad dollars. It's also twisted discourse to be sour and negative by default because it activates these perverse reward pathways.
Do you only buy hand-churned butter?
Most fliers put together for community events pre-AI were Microsoft Word or Google Docs plus clip art.
The world is getting better, but before everyone realizes it we have to complain first and reminisce about the good old days of Wingdings and Comic Sans.
Some of them would have passed as intentionally crafted by a graphics designer before AI.
And let's not forget that before AI, many of these would be word doc printouts with clip art. The alternative wasn't hiring a designer - the alternative was doing nothing special at all.
Google needs to be broken up.
Universities are great for networking, starting projects with other students (not the ones professors mandate), and learning lab sciences. In research, they're great for institutional knowledge, having a community of peers, getting guidance from research advisors, having real equipment and funding, etc. But there's a great need for AI tools to accelerate learning outside of that setting.
Anecdotally, I'm a working adult. I'm not going to waste time in college again. I need this for me.
Do you think people will stop dating, trying to impress mates, buying luxury, etc.? Neither candlelight dinner to a robot no a Dior manufactured by robots have the same allure. There's a huge industry around this.
Sports aren't going to go away. Huge industry.
People aren't going to stop making art. I know a ton of artists who have embraced AI that are doing even bolder work using the tools. (I was a filmmaker pre-AI, and I know a lot of people in this field.)
People aren't going to stop traveling. And consuming. And eating human food and consuming human experiences.
There are going to be all new kinds of businesses and opportunities that spring up. OpenAI and Anthropic are not going to be the only two employees. They won't be staffed by only agents.
I'm hoping someone develops an interactive tutor that can teach any subject to any depth.
The tutor should optimize its pedagogy. It should use online RL to adapt to a learner's ideal learning style, model what the student understands and to what degree, and understand what the gaps and next steps are.
I'd subscribe in a heartbeat.
The cost will tend towards zero because competition is intense and there appear to be zero moats and ample improvements from every direction.
If these systems have low costs, that means they will be usable by the broad population and that their utility will be widely accessible.
Industrialization led to iPhone, PlayStation, Spotify, and Waymo.
AI will lead to personal chefs, contractors, assistants, drivers, tutors, climbing partners, ...
AI will lead to people making their own PlayStation games, their own music and streaming services (I already have), their own custom smartphones (a future personal project - vibe hardware). And your robot will drive you cross country on vacation while you sleep in the car.
[1] Not actually infinity because earth [2] has finite resources, but the S-curve will look like it for awhile.
[2] Until the robots leave earth, anyway
Democracy is important and it isn't free.
That will be a decacorn product or more.
We should have web installs by now. The only reason we don't is because Google and Apple like cash and their little monopolies are easy money.
Big tech loves to "protect us". See Anthropic and OpenAI worried about intelligence.
Google doesn't care that its AdSense ads marketplace is flooded with malware. Or that YouTube is rife with scams. Wonder why not. The blatant policy contradiction couldn't be because money, right?
The labs frequently apply their raw models to problems that do not make economic sense for their customers but that demonstrate the power and capability of their systems. These experiments can cost millions of dollars. That's not customer-shaped.
They're not going to give you access to that. It's not a product. The government might have an interest in this, but that's not something you'd be privileged to know about.
And when these labs do develop "AGI", they more than likely won't be selling it to end users. They've pretty much already said this.
Y'all seem to miss the point of this forum. Building and hacking and science and engineering.
I swear there's a whole lot of you who just like to look down instead of up. There's a whole universe up there.
People are holding it wrong, deliberately or not. Some are inventing bad faith measures so they can claim AI sucks.
You're thinking about this the wrong way. The system is built and delivered as it is because that's how the providers make the most money. If they cared to have it perform well in chess games, you'd see a different shape and behavior.
We shouldn't ask the multibillion dollar automated software generation system to play games with us any more than we should ask a Boeing's flight guidance system to do so.
One way to look at this is that the internet is healing. Platforms are closing themselves off to growth and scabbing over.
I'm not a fan of Musk, but his driverless cars are already driving around in my city alongside Waymo. His rocketship thing will probably be delivering tanks to the other side of the world within hours if the DoD gets their way.
I'm getting more work done every day than in entire months pre-2026, and I've done my share of hard engineering. I think you're the one coping.
They just solved a Millennium Problem for fuck's sake.