mg

31 ms·

571.
by mg 3y ago
I remember my first experiences with touch screen phones. It was magic, and it was clear that this would change the world. Touch screen phones were the biggest and final "dent in the universe" Jobs made. But it was not as magic an
572.
Free DALL-E 3 access sends Microsoft Bing user numbers through the roof(twitter.com)
1 points by mg 3y ago|0 comments
573.
by mg 3y ago
It's interesting, that Google keeps publishing AI research papers. Is there a business rationale behind it? OpenAI has become one of the fastest growing companies of all time. And much of it is based on Google's "Attention is
574.
by mg 3y ago
What makes you "very sure"? Look at the chart for bing.com: https://trends.google.com/trends/explore?date=today%205-y&q=... Do you attribute that to Chandler Bing too?
575.
by mg 3y ago
I've been watching the interest in ChatGPT and its rivals via Google Trends for a while: https://twitter.com/marekgibney/status/1724712305894600929 It's fascinating that ChatGPT seems to stand a chance t
576.
by mg 3y ago
Then let's see who manages to write the least interesting comment in this thread.
577.
by mg 3y ago
Well, that is about the question which is the most uninteresting number. And the paradox that - whichever number it is - this attribute makes it interesting. But since we are not looking for the most uninteresting number, but the most inte
578.
by mg 3y ago
I agree that this makes 37 somewhat interesting. Certainly more interesting than - say - 31. 31 is a prime number too, and therefore somewhat interesting. But for sure not as interesting as 37, which as we just learned, is the median value
579.
by mg 3y ago
A friend and I had a discussion today, whether Amazon was a small internet company among many, back when it IPOed in 1997. Or if it already stood out as one of the biggest IPOs of that time. To our surprise, we did not find a list of dotcom
580.
by mg 3y ago
You don't need to use an open source LLM for the approach I described. You can still send the prompts to OpenAI's GPT-4 or any other LLM which is available as a service.
581.
by mg 3y ago
So these "GPTs" are the combination of predefined prompts and custom API access? Not customly trained LLMs? If so, I guess you can make such a "GPT" on your own server and independent from a specific LLM service by using
582.
by mg 3y ago
Well, since html elements already have a querySelectorAll function, you could just assign it to a new method "qsA": HTMLElement.prototype.qsA = HTMLElement.prototype.querySelectorAll; Now myTable.qsA('tr') will
583.
by mg 3y ago
import {qsA} from './dqs.js'; rows = qsA(myTable, 'tr');
584.
by mg 3y ago
handlebars.js runs in the browser: https://plnkr.co/edit/LXb5TOMoyAxdbTxB
585.
by mg 3y ago
That's what I did before I put those lines (and some more) into dqs.js
586.
by mg 3y ago
What's wrong with solo devs? You are talking to me on a site done by one. This is the frontend code: https://news.ycombinator.com/hn.js
587.
by mg 3y ago
These days, I get by with just two dependencies. One is dqs.js: https://github.com/no-gravity/dqs.js Which turns document.querySelector('.some .thing') into dqs('.some .thing'). The other one is Han
588.
by mg 3y ago
For humans, it works.
589.
by mg 3y ago
In wonder if copyrighted content will be needed at all in the future of AI training. AlphaZero learned to play chess via self-play, not by reading books about chess. Why couldn't the same happen for art for example? For coding, won&#x2
590.
by mg 3y ago
Is there no type of work a human can do in a short time which has value? Say in 20 seconds. If there is such a task, then it can't be automated. Because then it would not have value. So we could base a captcha on performing that task.
591.
by mg 3y ago
Apple claims the transition from TSMC's N5 to N3 has resulted in a 30% boost in performance for the M3's performance cores compared to M1's. Is that 30% more performance while weight and power consumption ar
592.
by mg 3y ago
There are more nuclear warheads in Russia than in the USA.
593.
by mg 3y ago
The moat he describes the one-percenters might aquire is government regulation. Creating an environment where only an exclusive club of companies is able to work on AI. Is that possible? Has that happened in any other industry? I can imagin
594.
by mg 3y ago
This is approaching the code size of the vanilla js version. And it does not have the confirmation modal. If there would be additional computational logic involved (say calculating how many rides this car has done this year so far to say &q
595.
by mg 3y ago
Hmm.. I don't think you can simplify their example to just that one line. For example, where is the url which is supposed to be requested to delete the item on the server.
596.
by mg 3y ago
I wonder what to think of htmx. Their primary example is this: <button hx-post="/clicked" hx-swap="outerHTML"> Click Me </button> But that is rarely what I want to do when a but
597.
by mg 3y ago
In Vim, I tried to assign a shortcut to send the selected text to Phind (or any other LLM) and came up with this: :'<,'>y|call system('firefox <url>?q='.shellescape(@*).' &') The only problem le
598.
by mg 3y ago
Since there is no demo and the repo looks super barebones (a loop which templates a list of mp3s in a directory?), do the upvotes mean that there is demand for an HTML based audio player? What features would everybody want from such a playe
599.
by mg 3y ago
It is interesting, that the lack of a feature that would cost $629 to add is significant enough to make it the the HN front page. Makes me feel like I would like to learn more about open source. What drives it's development and what th
600.
by mg 3y ago
I dislike what Musk has done to Twitter so far. It feels closed, greedy and messy. But in fairness, one has to say that after building a $650B company (Tesla) and a $150B company (Space X), it is not necessarily a sign of business failure w

More ›