Viewing profile — aurareturn
aurareturn
HN member- Joined
- Mon, Nov 14, 2022, 10:45 AM UTC
- HN karma
- 7,849
- Public activity
- 3,743 items
- HN profile
- View on Hacker News ↗
About aurareturn
Recent public activity
-
comment
Comment #49220040
It's far better to have the LLM crown because the crown is what will give you the highest margins.
-
comment
Comment #49210773
Yep. Called it: https://news.ycombinator.com/item?id=48033801 I've been calling it for years. The "AI has no moat" people on HN who keep repeating that singular Google employee's e…
-
comment
Comment #49207283
So the plan is to hide in this area in case of a nuclear war because countries won't target that area. Well, they just announced to the world that very important people will hide t…
-
comment
Comment #49207137
I don't see how being faster at doing something is worse.
-
comment
Comment #49207027
So the article argues that misdirection of the game is the reason budgets balloon and that AI won't help. I disagree. I think AI does help because it allows you to prototype far qu…
-
comment
Comment #49205794
Did Google cause Facebook?
-
comment
Comment #49197810
keeping Search profitable without enshittifying it I don't think Google had much to do with it. It has way more to do with the internet concentrating into major communities where a…
-
comment
Comment #49196398
Ok, so they're using footnotes. Can you now analyze backlog? For example, Microsoft and Amazon have backlogs of $1.45 trillion. Not including Google, Oracle, Meta.
-
comment
Comment #49178389
The winner isn't going to take all. What does take all even means? That the winner will develop the smartest AI, hack governments and launch nukes to eradicate the enemy completely…
-
comment
Comment #49177638
Did you read your own article? It's well less than $3 trillion. Now let's build the model out more. What is the projected revenue, backlog, improvements in existing big tech busine…
-
comment
Comment #49168717
Can we see some actual numbers, projections, models instead of vibes?
-
comment
Comment #49167958
Well, if no body can say when it will pop, then can we really say it's a bubble and it's overvalued? I can tell you it will pop in 10 years and when it pops, it will still be 20x b…
-
comment
Comment #49167197
When is it popping? Is the AI bubble in the room with us now?
-
comment
Comment #49163565
Honestly, so much illegal stuff happen on Telegram, I'm surprised they didn't remove it earlier.
-
comment
Comment #49163539
A lot of HN posters thought it was already bending at GPT4o/GPT4.5. Turns out, it kept accelerating (at agentic tasks).
-
comment
Comment #49155699
My brain didn't work there.
-
comment
Comment #49152050
I'm pretty sure you can adjust the creativity for many Chinese model inference providers.
-
comment
Comment #49151523
It'll sell after 1-2 times, then the market will get jaded and ignore.
-
comment
Comment #49151469
Check their R&D expenditures and profit margins they are quite poor. They're not. They're some of the top companies in America by revenue, profit, and profit margin.
-
comment
Comment #49151345
Lmao I love their video with the idea that people will be able to do their hobbies while ai does their job. Anecdotally, I'm even more busy with AI than before AI. I'm expected to …
-
comment
Comment #49151324
You will simply not get more value out of running a local model vs paying for a subscription/API from the cloud in 2026. There is no math that will make local models come out ahead…
-
comment
Comment #49140966
they copy the products with lower margins then western companies can't compete on price. I think this is outdated thinking in 2026. It's more like western companies go to China, lo…
-
comment
Comment #49121752
No one said Fable 5 is human. That's like saying Fable 5 should have told me how to prompt it to solve the Riemann hypothesis like Terence Tao would. The skill of the user still ma…
-
comment
Comment #49121736
Then start your company and go compete against OpenAI and Anthropic then. Heck, just take Kimi K3 and start your training run from that.
-
comment
Comment #49121708
You're using it wrong. No really you are. Give Fabe 5 access to a test database with some data, or even restricted access to your live DB and tell it to optimize then. I've had stu…