This is actually quite impressive from an engineer viewpoint. I just have the feeling that the videos quickly become very monotonous and rather boring due to the monotonous AI voices. If somehow you could bring dynamic variation in these videos that would be fantastic.
Presumably you also like that they are explaining things right? I mean that seems like the more critical step. Otherwise if it's just Marquise Brownlee, you could just watch Marquise Brownlee say the same sequence of random words for 3 minutes a few times per day.
I think this is exactly the opposite of what AI should be used for. It is going to make people dumb.
Watching a video instead of engaging your own brain makes you feel like you learned something without actually trying and I would bet it works much less well.
If someone is there adding something beyond what is already there - ie someone like Maruqise. Then it makes sense for them to be there.
If not it is just brain rot.
People already have issues with concentration. Allowing them to further allow that muscle to atrophy will not be good.
All of what you said is great, and I especially detest the proliferation of slop videos on YouTube, where it makes no sense to drown out the ample supply of such personalities and insights with AI voices summarizing wikipedia or Reddit threads over AI imagery slideshows.
But given how good a job this seemed to do at giving me more than just headlines, I'd love to have maybe even just an audio podcast feed of the top 10 stories like this compiled a few times a day. I would listen while I'm doing things when reading isn't practical.
tl;dr agree that we don't need this to replace reading, but I see that it can be a useful tool.
Yeah, there’s been an explosion of that guff over the past few weeks. “POV: You are a $x who $y”. “Every level of being a $x”. “$x explainer”.
The quality is getting very smooth but that just makes it worse because you can’t quite trust anything any more. It’s like the mental equivalent of HFCS.
I just had opus make me a tts mobile app that grabs the top 20 HN stories and the top 5 comments for each and read them out to me (using one of the higher quality built in TTS voices on my pixel). The summarization is the worst part of this imo, so pure TTS wins
Yeah I do see your point. And for some reason making it into audio does seem less offensive to me. Maybe it’s because it at least transforms it into something I can do while doing something else. Whereas a video is basically the same mode of operation: staring at a screen.
I’m still basically against this though. I consider it slop.
Thanks! And I agree. We're getting tons of requests for more nuance in voice selection from our users. You're currently able to say i.e. "Australian English female" in your prompts today, but you should also ideally be able to describe the voice characteristic (i.e. like a funny grandpa, engaged news reporter).
What Gemini 3.8 Flash TTS is doing with generative voice design in this area super interesting.
As a chess player I feel this is fascinating although there are plenty of analysis options already on lichess on chess.com. Still, sometimes it is more about learning and curiosity. Well done.
After the "metaverse" hype of 5 years ago, VR clearly hasn’t really taken off.
I’ve personally always thought the headsets were simply too clunky, which puts off the average person from trying them, due to the hassle (and price).
This new design from Meta is the first I’ve seen that feels like something more people would actually want to try. Perhaps VR isn’t dead — maybe it’s just lurking in the shadows, waiting for the right hardware tech to bring it to the masses.
I am starting to wonder how useful these benchmark still are? Aren't all of these models trained to ace these benchmarks?
In any case they give an indication, but I am increasingly looking to real world feedback from real users. I have a solo project I (voice-to-text typing voicewink.app) and I'm using both the Claude Code and Codex coding harnesses and multiple agents doing reviews on the same code.
This has given me real tangible results to compare on real work. Conclusion: GPT-6, Opus 5 / Fable and Grok are "top tier", with Claude good at planning and executing and Codex / Gpt-6 better at finding bugs and fixing them (but tending to overengineer), and Gemini and Kimi 3 clearly behind in capability (more so than the benchmarks suggest in my opinion).
There is something weirdly fascinating about this. I think the combination of nostalgia, seeing others play without interacting and just the whole idea of it. Rather mesmerizing. Well done.
https://docs.mistral.ai/inference/model-selection-guide?mode...
Cost is stated at half the price of GLM-5.3, which is quite interesting.
reply