throwaway4aday

7 ms·

1.
by throwaway4aday 1y ago
Why not just write a script that does this but with all of the model providers and requests multiple completions from each? Why have a whole ass editor open just for code review?
2.
by throwaway4aday 1y ago
I think you need to be more specific about which "chatgpt reasoning model" you used. Even the free version of chatgpt has reasoning/thinking now but there are also o1-mini, o1, o1-pro, o3-mini, o3, and o4-mini and they all ha
3.
by throwaway4aday 1y ago
They only did that for image generation. The more interesting part is that an LLM can approach or find the correct caption for an image, video or audio during test time with no training using only the score as a guide. It's essentially
4.
by throwaway4aday 2y ago
Are you contending that every human derives their reasoning from first principals rather than being taught rules in a natural language?
5.
by throwaway4aday 2y ago
Doing better than Firefox
6.
by throwaway4aday 2y ago
It seems that you didn't understand the main point of the exposition. I'll summarize the ops comment a bit further. Points 1 and 2 only explain how they are able to erroneously justify their absurd beliefs, they don't explain
7.
by throwaway4aday 2y ago
That's essentially what R1 Zero is showing: > Notably, it is the first open research to validate that reasoning capabilities of LLMs can be incentivized purely through RL, without the need for SFT.
8.
by throwaway4aday 2y ago
Worked great for Vancouver /s
9.
by throwaway4aday 2y ago
This is a quick "fix" that has a lot of unintended consequences. I've seen it first hand and the only people it benefits are those who are already wealthy. Everyone else gets a lot poorer as their cost of living skyrockets. H
10.
by throwaway4aday 2y ago
I'm not sure if you've read many comments on here but there seem to be more people railing against generative AI than there are those sharing useful and interesting stories or critiques. It's one of the reasons I've been
11.
by throwaway4aday 2y ago
Step 1: https://claude.ai Step 2: Write out your description of the thing you want to the best of your ability but phrase it as "I would like X, could you please help me better define X by asking me a series of clarifying q
12.
by throwaway4aday 2y ago
Not so sure about those examples and pairing with the idea of quick and dirty work.
13.
by throwaway4aday 2y ago
Your point is on the verge of nullification with the rapid improvement and adoption of autonomous drones don't you think?
14.
by throwaway4aday 2y ago
I don't think motor skills are a good object to use in an argument about verbal vs non-verbal thinking. We have large regions of our brains primarily dedicated to motor skills and you can't argue that humans are any more talented
15.
by throwaway4aday 2y ago
Yes! that too.
16.
by throwaway4aday 2y ago
The tower is also how they plan to perform maintenance and re-staging for another launch. The tower can place it back on the launch structure or lower it down to the ground if it needs to be transported back to an enclosure for extensive re
17.
by throwaway4aday 2y ago
true but don't forget this is test flight #5
18.
by throwaway4aday 2y ago
not sure how anyone can miss the "throwing away the airplane" part as being the cost driver.
19.
by throwaway4aday 2y ago
With a payload volume of 8m diameter by 22m height you could fit a James Webb size telescope inside with minimal folding. The sunshield (21.2 m by 14.2m) would only need to fold along one axis and the mirror (6.6 m) could be monolithic inst
20.
by throwaway4aday 2y ago
You're pretty good at writing clickbait headlines yourself. Those huge checks are a DoD contract for unblockable internet coms not some handout and for $20 million a month it sounds like Elon is giving them a deal, at least compared to
21.
by throwaway4aday 2y ago
Unfortunately, they've taken to doing things like saying Horse C is in the lead when it's actually Horse B and telling everyone Horse A has gone lame when it's clearly neck in neck with Horse B while pretending that Horse D d
22.
by throwaway4aday 2y ago
someone should revive it as a GUI debugger. doing so would narrow the scope which should cut down on complexity. let people use their own editors and IDEs for making code changes. the focus should be on inspecting state, seeing live updates
23.
by throwaway4aday 2y ago
Wow, great work and kudos for making the source available! I'll have to give this a try.
24.
by throwaway4aday 2y ago
we're still very far away from complete automation. one person now can do the work of 10 or 100 workmen in the past but people are still needed to transport and set up the feed stock, monitor, configure and service the machines, remove
25.
by throwaway4aday 2y ago
I don't think there's a single way that we learn things, there's too much variety in how, when and why things are committed to memory and still more of a difference with things that actually update our thinking process or wor
26.
by throwaway4aday 2y ago
> it's pretty jarring if anyone can hear it That's only because you haven't put any effort into sound design. Notification and alert sounds can be horrible like many of the ringtones people choose or they can be pleasant
27.
by throwaway4aday 2y ago
I would think they'd be more likely to form the language centre of a composite AGI brain. If you read through the known functions of the various areas involved in language[0] they seem to map quite well to the capabilities of transform
28.
by throwaway4aday 2y ago
Setting aside land and limiting or not allowing development of it so that it can be publicly available for enjoyment of nature or hunting and fishing is a laudable endeavour. Banning all human presence and use of large swaths of land, on th
29.
by throwaway4aday 2y ago
I didn't say anything about a timeline for _solving_ these, just that the short timeline for drop off in demand for compute is unfounded since there is still so much ground to cover. The article takes the shortsighted view that the cur
30.
by throwaway4aday 2y ago
While I agree that we don't want to extrapolate too much I disagree that this type of exploration may not benefit from more compute. We won't know until we try and since we have what seems to be a very generalizable architecture i

More ›