throwuwu

23 ms·

241.
by throwuwu 3y ago
You can write bad software for any system. Their plug in store needs a lot more features for reviewing and identifying what a plug in does. A friendly interface for the yaml file would go a long way. No sane person would enable this plug in
242.
by throwuwu 3y ago
Well written plugins already do that. This one is not well written. It provides a single function call that can run any octokit command with no confirmation from the user. Foot, meet gun.
243.
by throwuwu 3y ago
Better yet: crappy plug in that executes arbitrary commands with no human confirmation allows GPT-4 to execute an arbitrary command with no human confirmation.
244.
by throwuwu 3y ago
jfc the plug-in is just a blank check. Here’s another bright idea: eval-gpt-plugin
245.
by throwuwu 3y ago
It’s hard to tell why you ran into such a problem without seeing how you prompted but I can offer a few pointers. Use the OpenAI playground instead of chat, it allows you to specify the system prompt and edit the conversation before each su
246.
by throwuwu 3y ago
It’s crazy how many people miss this. GPT models can review code too! They can also write and run tests. Once the context window is big enough to fit the whole code base into it they will be better at review than you are. Eventually we’ll h
247.
by throwuwu 3y ago
If gpt writes it and I don’t understand some parts I just paste it into a regex testing site and examine the groups and try it out with different snippets of text. Is that so hard? Is it any different from using something you found on SO?
248.
by throwuwu 3y ago
For me it saves time by keeping my momentum up. I don’t lean on it when I’m in a flow state and cruising through the code I’m writing but as soon as I hit a wall I jump over to chat and start working on a solution with it. This saves a huge
249.
by throwuwu 3y ago
Most of the complex regexes I write fall into the category of stuff I don’t understand that might work. I’m no slouch at regexes either but when you start trying to do weird data processing stuff that handles all kinds of edge cases the whe
250.
by throwuwu 3y ago
At some point the models will produce code with a lower error rate than existing libraries.
251.
by throwuwu 3y ago
The only way to avoid this is to make the business a hobby too.
252.
by throwuwu 3y ago
wdym by latency? From what I can find the vision pro is 90hz which is the minimum for sota headsets with 120 being targeted by higher end devices. There are lots of hand tracking solutions for headsets that don’t have it built in but for a
253.
by throwuwu 3y ago
My understanding is that Microsoft research has already published a paper where they used synthetic chat interactions of the same form that chatGPT uses to train a new model. GPT4 could be used to select the best interactions from which to
254.
by throwuwu 3y ago
While it can act as a mirror it is not only a mirror. There are many strategies that work with it e.g. bedtime stories, emergency, post apocalyptic, role playing, encoding, leading by example, etc. You can steer the probabilities and get ar
255.
by throwuwu 3y ago
1. Not having to sit at your desk where your monitor is. You can switch to sitting on a couch or in a recliner or walking on a treadmill or on your porch. 2. Taking it with you easily. You have the same large screen (s) with you in the hote
256.
by throwuwu 3y ago
There are other high resolution headsets on the market right now at comparable or less cost than the vision pro.
257.
by throwuwu 3y ago
Maybe you should ask chatGPT to explain what a cargo cult is
258.
by throwuwu 3y ago
The point is they don’t need any code. They can copy paste a csv or something even messier and ask it to write a somewhat personalized email to each person in the list, if they’re smart about it they can include a few small details with eac
259.
by throwuwu 3y ago
wtf? Who is this true for?
260.
by throwuwu 3y ago
I’m not even remotely concerned about an AI bubble, in fact the faster it can inflate and pop the better. An AI hype winter would be as comfortable as a tropical vacation with the current tech that’s available. We could build and research i
261.
by throwuwu 3y ago
Proposal: prefix your answer with GPT-3.5 if you’ve only used the free version and GPT-4 if you use the paid version.
262.
by throwuwu 3y ago
I don’t even bother with blog posts anymore unless they are new information and in that case it’s often easier to use Edge and Bing chat to interrogate the page directly. It’s far better than skimming or ctrl-f
263.
by throwuwu 3y ago
The original quote mentions a fence or gate built across a road. Also Chesterton is writing from an English perspective which includes familiarity with manmade landscapes built up over thousands of years where the intentions of the original
264.
by throwuwu 3y ago
Now that I’m experimenting with blocking YouTube altogether for longer and longer periods of time, I find that I don’t need it at all anymore. I used to heavily rely on it for tutorials but ChatGPT has replaced that; why watch a 30 minute e
265.
by throwuwu 3y ago
You conveniently left out the media hype cycles
266.
by throwuwu 3y ago
This is a bad interpretation of the argument. Obviously they meant that other creatures can feed themselves not that they derived their nourishment from light and raw chemicals. Asexual reproduction does not require a partner by definition
267.
by throwuwu 3y ago
Appeal to authority isn’t interesting. Tegmark is a physicist and science communicator anyways so he doesn’t belong in your list.
268.
by throwuwu 3y ago
One is the world’s dominant gateway to the internet, the others are a handful of websites.
269.
by throwuwu 3y ago
Humans aren’t robots. Just because you pay them doesn’t mean they’ll do anything you say, they have to make a choice as to whether they’re alright with doing your dirty work. It’s not like you can just rent mercenaries off of Craigslist and
270.
by throwuwu 3y ago
He still rules by consent, if the rest of his government decide they don’t like his style anymore then they can get rid of him. The original comment was about fully autonomous AI in the context of software making hardware making software wh

More ›