Pretty sure the references you are referring to are about losing coding skills, which is a very different set of skills than language skills. The kagi link a sibling comment left was an example of an AI that can improve writing while also informing about better writing style. As opposed to the slop grenade, which is outsourcing thinking to AI. The point is, you don't have to outsource your thinking to use AI, if it's a good AI tool. Problem is, most AI companies are coming from a hyper-scaling mindset where addicting the user to the product is the same as substituting hard thinking for easy dopamine hits. The most ridiculous benchmark I have seen in AI is the tendency to say the longer an agent can work and some minimal accuracy is a good thing. That just means you have 30 steps to find an error in instead of 3, and you are much more likely to just abdicate thought instead of the hard work of proofing 30 steps yourself. AI companies and evaluators are losing the point.
I wouldn't worry too much about it. I think the big AI standout (not your fault) is the classic GenAI lightning that behaves more like a weird viscous liquid. Another one I see a lot is the number of avian creatures flapping their wings with all the subtlety of an ornithopter going in reverse.
Feedback
Your site needs some work on the responsiveness. At around 1200 pixels width on my browser, half of the video just gets cut off. It just doesn't scale correctly.
You can get it to reproduce content but it’s a game of cat and mouse. Were it not for the alignment to avoid direct reproduction it would taken far more often.
Heck no. The world would be a way better place with no personal automobiles. Trains, yes. Even trucks and buses, sure. Cars, nooooooooooo. Cars are among the most clearly net-negative inventions to come out of industrialization. They should be criticized and fought until finally defeated. Self-driving cars are a massive waste of human and physical resources to provide a solution that is still strictly worse than proper urbanism and transportation network design.
I (and I assume many others on especially HN) don't consider intellectual "property" as real so there is no "theft" in our minds, so this argument doesn't bear much weight.
> As far as I can see people always radically exaggerate the effect of the incompleteness theorems
Like people saying Godel theorems "prove" LLMs could never "invent" new mathematics because being a software system Godel applies to their operation, but not to humans which are not math.
Or, just produce a cheaper, shittier widget, and pay Google to have their AI tell people your widget is superior anyway. My guess is Google will try to keep their ad costs just low enough to make this second option the most attractive.
My answer is if you want to code as a hobby -- yes; if you want to code as a job -- no. The reason I say this isn't that corporations won't still need some people with coding skills even in a world where all or most code is generated by AI, it's that the number of jobs is such that there isn't room for new people to enter the industry. Many jobs are being lost and few new jobs are being created. A new person could focus their efforts on something else, such as an apprenticeship to become an electrician, which would be much more likely to pay off and much more lucrative. It looks like manual dexterity will be the hardest thing for AI to automate, so a job that combines human cognitive skill with manual dexterity, such as being an electrician, is probably a much smarter way to go than learning to code, if one's goal is making money. Do you all think I am out of line in saying this?
I'm inclined to agree with Casey Muratori who suggests we think of coding like playing guitar. Should you learn to play the guitar in 2026? Very few people can make a living playing guitar. It's not impossible -- even in the age of synthesized guitars and now AI generated music, there are still superstar rock stars who make a living playing guitar -- but it's a very small fraction of people who can play the guitar. If you want to make a living playing guitar, then you probably shouldn't learn it. On the other hand, if you want to play the guitar for its own sake, as a hobby, then yes, you should learn it.
Could a similar argument be made about your comment, perhaps? :)
I wasn't being intentionally pedantic. I was in fact making a point that the reality will be a lot more grim than watching a giant fireball turn into a mushroom-shaped cloud for a few seconds or minutes.
That's almost certainly illegal in many jurisdictions, and they'd definitely not be able to hide that they're doing it indefinitely. A sure way to be massively sued.
Their open weight on device models are really impressive. Partly because I think they are the only ones out of all the frontier labs even working on local models.
We should instaban anyone who uses these patterns on ycombinator. If you talk like that for realsies, well too bad, try being more human and less linkedin next time
I use the CLI agents (from any major vendor), in conjunction with either nvim or standard VS Code (with Copilot disabled). That way you still get the automatic "agent" capabilities - it can search your code, propose and make changes, write tests, doc files, etc. - but it doesn't interfere with your editing experience.
The only thing worst than a mega corp is an ip attorney.
Your cause is already lost.
Good luck enforcing whatever frivolous lawsuits you have cooking up against open weights Chinese models that anyone with newer graphics card can crank out inference on.
> Highlighted Answers: When people are researching, they want helpful suggestions. Now, when AI Mode provides a list of recommendations — like the best language apps for an upcoming trip — highly relevant, high-quality ads are eligible to appear on that list as a Highlighted Answer.
The user asked for the "best language apps for an upcoming trip"?
Are you going to answer their question objectively?
What if the correct answer is apps A and B are what they need, yet the publisher of app C paid you to be a Highlighted Answer?
What if C is not only not among the best, but is a toxic load of poo?
Also, when are you going to stop often blatantly plagiarizing things people wrote on the Internet, for your "AI" answer, even though you absolutely know you're blatantly violating the social contract that built your company, and destroying the creators from whom you're stealing?
In end user perspective, it’s the same. The difference is probably volume. I have no clue in which direction. Both in trained lies by models and in number of people defending the indefensible. But one for sure, there are probably 10s or 100s of millions of Chinese who try to do the same. I encountered with it quite frequently. Sometimes with flat out doublespeak.
