Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ivape
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
24 ms
·
511.
▲
by
ivape
1y ago
Their beefing over South Park as well: https://www.hollywoodreporter.com/business/business-news/sou... At the heart of the dispute: a new 10-year, $3 billion overall deal for Parker and Stone that would more than
512.
▲
by
ivape
1y ago
I just bought a monitor arm that has a USB dock at the base. The dock is actually just a USB extension cord that is enclosed by the base of the arm. Think of it like putting a $10 USB hub into a box and poking holes in it to stick wires thr
513.
▲
by
ivape
1y ago
Everyone that successfully avoided social media for the last decade escaped with their mental health. Everyone that carefully moderates their ai intake (e.g don’t depend on Claude Code) will also escape with their skills over the next decad
514.
▲
by
ivape
1y ago
What’s the cost of pointing it to Private Cloud Compute? It can’t be free, can it?
515.
▲
by
ivape
1y ago
”The people who YOLO it with prompting cycles until the code passes tests and then submit a PR are causing problems almost as fast as they’re developing new features in non-trivial codebases.” This might as well be the new definition of “
516.
▲
by
ivape
1y ago
Slower in initial delivery maybe, but the maintenance and debugging of production applications requires intimate knowledge of the code base usually. The amount of code AI writes will require AI itself to manage it since no human would inund
517.
▲
by
ivape
1y ago
We’re also headed into a world where there will be very few open weight models coming out (Meta going closed source, not releasing Behemoth). This era of constant model releases may be over before it even started. Gratitude definitely needs
518.
▲
by
ivape
1y ago
I would say you were less of a douchebag for taking those gigs once upon a time than now days.
519.
▲
by
ivape
1y ago
What is the perfect amount of money and time for you? I suspect many things are not good enough for you and others.
520.
▲
by
ivape
1y ago
I don’t want the visual clutter.
521.
▲
by
ivape
1y ago
Maybe it has something to do with the “you are not good enough” treadmill the modern world has everyone on. I don’t think people of yesteryear contemplated if they were ready to start a family. I don’t think they contemplated if a job was t
522.
▲
by
ivape
1y ago
OpenAI hides their model's CoT (inference-time compute, thinking) Probably because Deepseek trained a student model off their frontier model.
523.
▲
by
ivape
1y ago
Someone else can confirm, but from my understanding, no they did not know sentiment analysis, reasoning, few shot learning, chain of thought, etc would emerge at scale. Sentiment analysis was one of the first things they noticed a scaled up
524.
▲
by
ivape
1y ago
"but what if it's true?" There was nothing hypothesized about next-token prediction and emergent properties (they didn't know scale would allow it to generalize for sure). What if it's true is part of LLMs story
525.
▲
by
ivape
1y ago
The DoD, DARPA, NASA, tax funded arms of the US should have all the resources to train a frontier model.
526.
▲
by
ivape
1y ago
It’s easiest way to repay Elon’s $150 million. Now listen to this song again as you digest that: https://youtu.be/VYOjWnS4cMY
527.
▲
by
ivape
1y ago
It’s more about not getting embarrassed ever again. The breakthrough papers for LLMs came from Google, but not the product. This was embarrassing. Even if you have to spend double on everything from now on you do it, because Google effectiv
528.
▲
by
ivape
1y ago
If we believe that there’s not enough hardware to meet demand, then one could argue this helps Apple meet demand, even if it’s just by a few percentage points.
529.
▲
by
ivape
1y ago
Personally? I find the damn thing to be magical. Truly. I can’t stop being amazed by it, so you’ll have to ask someone less head over heels with it. But I’ll grant you this, having said the above, I have no way of actually objectively provi
530.
▲
by
ivape
1y ago
GPT 3.5 was only released in late 2022. It really hasn’t even been a full three years. We’re still discovering what we even want to do with this stuff. Shocking when you say it out loud, it hasn’t even been three years . Let the kid grow u
531.
▲
by
ivape
1y ago
I’d also add that no one predicted the emergent properties of LLMs as they followed the scaling laws hypothesis. GPT showed all kinds of emergent stuff like reasoning/sentiment analysis when we went up an order of magnitude on the numb
532.
▲
by
ivape
1y ago
What could Google have possible offered? Must have been astronomical. Sundar has been aggressive ever since OpenAI knocked the wind out of them by being first release gpt.
533.
▲
by
ivape
1y ago
It's very fixable. Everything else is amazing, I love it.
534.
▲
by
ivape
1y ago
I'm speaking for its place in video games. Stuff like Hearts of Iron and Diplomacy are birthed from Risk but are extremely niche. In fact, for fully graphical video games, they often have board-game like communities (the games have no
535.
▲
by
ivape
1y ago
You need to make it clear what my starting position and color is.
536.
▲
by
ivape
1y ago
Risk Games are kind of niche but have a hardcore base. Check out Risk Legacy SC2 custom map, and Line Wars on Steam.
537.
▲
by
ivape
1y ago
The default is a remote Gemini call?
538.
▲
by
ivape
1y ago
What is the default BrowserOS model? Is it local, and if so, what inferencing server are you using?
539.
▲
by
ivape
1y ago
Still looking for an email provider with customer support I can call 24/7. I'm willing to pay for peace of mind.
540.
▲
by
ivape
1y ago
I was actually looking into reading some of his other candid works: https://en.wikipedia.org/wiki/Down_and_Out_in_Paris_and_Lond... Semi-autobiographical about when he was nearly homeless and living in poverty in Europ
More ›