HN user
Voice AI Engineer & Consultant
Interesting. I prefer video/audio for the same reason. But I watch most things on 1.5x or 1.8x speed. I'm not a particularly fast reader, so I prefer audio/video forms.
Totally agree with you though about not thinking written form is inferior. I do still enjoy a good long form article. (Also I'm not including technical writing like documentation.)
So should I ditch ollama for this?
Sincere question: have any non-Chinese models failed this one?
You telling me your kids don't want a bowling alley?
For just desktop, I use Handy (open source): https://handy.computer/
And yea, Nvidia's Parakeet v3 is good enough for my own just local transcription most of the time.
When I need local transcription to be more reliable and I don't have the energy to proof read a long ramble, I still often just pop open chatGPT, dictate, cut, paste.
But we're pretty much already to the point where local transcription models can replace cloud ones for personal use. They're still a bit rough around the edges in terms of polish and latency, but plenty of people are fine with that to avoid yet another app subscription and not having to worry about wondering what's potentially happening with their data.
You do know people aren't talking about fictional people being generated right? Actual, real-life, identifiable children.
https://www.theguardian.com/technology/2026/mar/16/lawsuit-e...
Very weird stance, to say the least, to not call that CSAM.
For me, it was the fact that grok was a CSAM spewing faucet for weeks. And also the mecha-hitler thing. And also that one time it was weirdly aggrandizing Elon.
I implement AI applications for enterprises (specifically in regulated sectors like healthcare and finance) and professional standards prevent me from ever recommending grok models. Way too much risk and liability for a business.
Too many times I have to say "groq with a 'Q'" just to make sure no one thinks I'm crazy.
But this isn't a thread about ice cream...? It's about SoTA emerging technologies that pretty much everyone agrees is going to have consequential effects on society.
When people say "don't talk about politics" that gets twisted up with partisan squabbles (like your example of bringing up the president during ice cream).
But talking about the quality of your streets, your local schools, your annoyance at the trash pickup service, or the data center in town that is using illegal gas generators and actively spewing methane around poor communities with little recourse...
All of that is "political".
Political means who gets to make decisions about how resources are allocated, at varying levels and scales of society. All the way down to the community level and all the way up. Who gets affected by the allocation of those resources. Where power is concentrated or distributed to or from.
Every thread on HN about open source, big tech, startups, etc. are invariably going to be political.
People caring about a trillionaire's outsized power in the tech industry and society at large is inextricably linked to the technology his companies create.
Yea but people invite actual humans into their homes who have names, faces, reputations, relationships, and some degree of social accountability.
If I hire someone to come into my home I can meet them, decide whether I trust them, build familiarity over time, and develop some form of reciprocity. They know whose home they’re entering, and I know who they are.
That feels very different from an anonymous person on the other side of a teleoperated robot... who may be one of many interchangeable operators, switching in and out on some unknown schedule, with no meaningful relationship to me.
Maybe I’m just the wrong audience for this. Because no way am I comfortable with anonymous strangers looking around inside my home.
Yea, the "doom trolling" seems to be starting to reach its limit.
https://www.nytimes.com/2026/06/17/opinion/ai-dangerous-open...
Hm, I just keep a folder called something like `status_docs/` in any project I work on and I create a new file in that folder any day I'm working on a project that's dated (e.g. `project/status_docs/2026_06_21_status.md`). It's basically a project diary that both me and the LLM can reference.
I have the LLM at some point in the day while working on the project create that file with all the relevant context. And then I'll periodically have it update that file (often before I compact the context window, or before I switch to a new task). And then I just have the LLM update it whenever I'm done working on the project for the day.
Then no matter what, if I come back that project again a day later, a week later, a month later, whatever – I just literally point a fresh session at the most recent status doc to help both me and the LLM orient ourselves to the work at hand. What's really nice too is having it reference the status docs from previous days to help orient it for creating the new status doc for the current day.
I've been doing this informally for probably over a year now, and have started formalizing it so I do it with every project. It's been a big help to me personally given all the context switching between projects I've been doing more and more since using AI coding tools.
"Dangerous" is a loaded term. But yes, even "soundwaves" can cause harm, same way use of pharmacological medical interventions can cause harm. Dosage, application methods, side effects, etc. all exist for medical use of ultrasound too. https://pmc.ncbi.nlm.nih.gov/articles/PMC8954895
Yup, precisely. Turns out getting AI to be reliable at doing useful things is harder than we've all been led to believe by the dominant narratives.
https://www.normaltech.ai/p/new-paper-towards-a-science-of-a...
Lets redefine work into discovering the world again. Let people do old handcraft jobs, let them do more sports, let them read more, let them write and make more. Let them enjoy nature.
Why leave something so important up to what AI does or doesn't do?
I certainly don’t think individual cells have any subjective experience or sense of agency.
There's definitely research and scholarship that would beg to disagree with you there. At least in terms of completely writing off the notion of "agency" when it comes to cells.
Dr. Michael Levin's lab is doing some pretty cool work. https://drmichaellevin.org/
something like the universal approximation theory comes to mind, transformer architecture clearly has the shape of a universal algorithm approximator
That says nothing about emulating a human brain.
I'm hearing different from PhDs. The bottleneck with much research isn't "trying out ideas" so much as it's all the bureaucratic minutiae, grants, mentoring PhD candidates, collaboration with other researchers, etc.
I've heard LLMs can be helpful in limited targeted ways. But not as some kind of "game changing" accelerant.
I call it "I'm not like other girls" writing.
4.8 is also 2x more expensive for a "modest" performance bump. How refreshing.
Where are you seeing it's 2x more expensive? https://platform.claude.com/docs/en/about-claude/pricing
I'd be kind of shocked if a model that came out six months ago is the same size and cost to run as one that just came out today.
Well, it seems like collectively we are all struggling to perceive model progress, given that it seems like every reply to you is reporting different experiences with which of the models has subjectively performed best for them.
I have my original gmail address from 20 years ago still and even old youtube videos my friends and I uploaded from ~18 years ago.
The cringe is rough but at some point the cringe becomes so bad it loops back around to me just feeling nostalgic and grateful that there's proof I was able to do things, create, be silly, whatever without worrying about appearances so much.
Also, I figure if I ever become a megalomaniac then old youtube videos of my teenage self doing parkour should go pretty far in humbling me (although, honestly, I think 13 year old me was way cooler than I am now, so I guess it could backfire).
Also it was but a few months ago that their CFO said, in a court filing, that Anthropic's revenue across the entire lifetime of the company "exceeds $5 billion". Pretty strange.
https://www.reuters.com/commentary/breakingviews/anthropic-g...
Yea, I wonder the same. Something I've been wanting to read up on is how the land ownership works with these data center deals.
Because developers promising massive projects to scoop up a bunch of land, do little to nothing with it, sit on it for awhile, and then eventually just sell it for a profit...that's not a new thing at all and isn't unique to the tech industry.
I wouldn't be surprised if a not-so-insignificant part of all of this winds up being just banking land that you got zoning, entitlements, and maybe some utility infra stuff setup for.
That may also explain some of the kind of puzzling cloak and dagger behavior with so many of these data center initiatives in local communities. If you truly think you're about to build something that is going to "imminently transform the way we do everything" and become some kind of $X trillion dollar industry, I'd imagine you'd be showing up with better "gifts" to ensure quick frictionless approval. But if they're more so viewing project proposals as speculative investments for control of land that is hopefully desirable to a bunch of tech firms (whether that's now or years from now), then keeping costs low and details vague early on makes more sense.
Yea, my bad. "majority getting scrapped” was sloppy wording.
More accurate to say would be that a big chunk of the AI data center pipeline looks delayed/speculative. 16GW is slated for 2026, but only 5GW is actually under construction. https://www.sightlineclimate.com/research/data-center-outloo...
I do think it's a bit ridiculous though to not consider someone a tech insider who was a director for a decade at one of the biggest tech companies in history...
I'd argue we don't need to build a 9 GW data center anywhere. But that's just me.
The data centers the industry wants are all going to get built.
If that was the case then why are the majority of data center projects getting scrapped?
https://gizmodo.com/data-center-project-cancellations-quadru...
Why are insiders saying they only expect about 10% of data center projects to ever be completed?
Why is 2026 already shaping up to have less than half of planned data centers break ground on construction?
Local community opposition is a big driver but so is permitting and infra procurement.
https://www.datacenterknowledge.com/energy-power-supply/why-...
All of this is inconvenient to big tech's "inevitability" narratives.
Heat. 9 GW puts off a staggering amount of heat. In an already arid environment. What's the worst that can happen?
https://www.sltrib.com/news/environment/2026/05/07/utahs-dat...
It seems fairly likely that a comically gargantuan data center like Stratos would endanger the surrounding ecosystem, at a minimum.
https://www.sltrib.com/news/environment/2026/05/07/utahs-dat...
But I also think there's very little chance that they even get 1 GW up and running anytime soon, and less than 3 GW before the whole thing gets scrapped (just like plenty of the other hyperscale data center projects that keep getting shouted about).
No, they do not have empathy or motivations. Arguably, if you think of them as having such then maybe it could help you coax out better outputs occasionally (wildly dependent on the task at hand). But that's only because of the LLM always wanting to "complete the story" -- "the story" being the prompt (which includes any "unseen" parts in the context window like a system prompt set by the application you're likely calling the LLM through).
It'd be more accurate to say that using language that tends to evoke empathetic motivated responses is more likely to get them. I'd argue that's only going to be relevant in scenarios where you want outputs that read as more... "empathetic and motivated".
The important point though is that none of the above equals "better" outputs, just different.