Occasionally in this case meaning every 48 hours or so, which means the limit is barely doing more than jack shit
922 karma · joined August 29, 2023
1: “safety”
2: proprietary protectionism and
3: it’s probably rapidly changing enough to be hard to pin down.
I am a massive proponent of this changing, it’s very strongly holding back these models. Change 1: models need to have more “LLM behavior analysis” in their training data/weight. Change 2: models need to have very detailed DETERMINISTIC logs of each step they take, and be able to access those logs. Change 3: the tuning and tweaking that happens more frequently needs to be in a .md file that the model can access.
Change 4: with the other changes done, the models should now engage in several self analysis steps layered into its whole thinking chain. “How did I reach this conclusion, did this require any guessing, do the key facts have research support online, quick check for Claudisms or AIisms and common LLM issues, did my changes alter underlying things like libraries without integrating them etc. etc.”
It’s how we think and refine our ideas and plans, and the models should mimic that.
Do programmers use this website anymore? Me, myself, I am a DOGSHIT amateur programmer and even I can tell these things are terrible without constant revision and oversight.
“Does the other side need to show any good faith?”
“No, in fact the judge is more likely to side either then the more angry and aggressive they are”
“….great”
The whole advantage of having any screen at all when folded is something quick to access and *one handable*. Just looking at the Duo, it appears too wide for that by a large amount.
If I’m forced to two hand the thing I’ll just get a phablet: no folding mechanism, no unfolding step before using properly, no gigantic price hike (guessing).
When we do meta thinking about that process after the fact, two things happen. 1, we change our total “thought” by adding that meta thinking to it. And 2: it’s a very lossy process, because we don’t have very good data about what our brain or mind was actually doing during that first think and emotional factors are nearly always at play and even more complex.
Now for the more complex AI, the fragmented process of multiple agents and loops and reruns are pretty similar to that first think we do. At least structurally. But the meta think is where they differ. They have no emotion, but they also have even worse data about its own function. They constantly degenerate so I would argue their “changing the thought by thinking about it” factor is also generally way higher than ours.
Getting better at consistent/reproducible thinking, with many ‘steps’, that leaves good documentation of that thinking behind for future analysis, has to be one of the more important areas for the big flagships going forward. I’m certain that “what is this fucker doing and why” is the biggest pain point for AI researchers. Or the math, it’s usually the math.
But you’re correct in the general structure; they generally do the same post hoc analysis we do, just noticeably worse because of their opaque nature(even to themselves) and general degenerative instability.
the inconsistency of LLMs is by far one of the biggest gripes I have with them. Closely related to their apparently deep desire to avoid following instructions.
I know these are both a byproduct of noise (which is somewhat tunable) and noise is inherent to these systems in a lode bearing way.
I still hate it. it’s holding the technology back. I don’t honestly see how we can safely or even successfully approach the idealized realm of AI without bypassing this problem, which to my understanding, probably means not using language models at all and trying a totally different approach. But I really don’t know much about machine learning, I’m a super novice compared to a lot on this website.
It’s why cancer is so prevalent and damn near unsolvable: an organism our size has about 7 octillion atoms, even forgetting the impacts of the stuff smaller than that. That’s 30ish trillion cells, many are multipurpose little motor vehicles inside our body running around doing stuff constantly.
It’s just so much STUFF happening, take the brain out of someone and leave them a vegetable digesting on a bed and they still burn hundreds of kcal a day, and that’s a highly efficient system millions of years in production.
So two things emerge from that, 1: extremely robust universal principles like electromagnetic gradients being used for travel, and 2: runaway feedback loops like black holes or cancerous cells.
What that means is that you can view mathematics as a spontaneous emergence FROM a high resolution system, or the inverse. We usually think of the inverse, because it seems like no matter what we discover in the system it can be explained by increasingly complex math.
But the inverse of that is very interesting to me: math emerges as a side effect from the high resolution system. It also passes the smell test, if you make a low resolution simulation of the solar system you will not see the effects of the high complexity systems, relativity works perfectly without any quantum dynamics.
I just love these kinds of chicken and egg problems.
Elected officials are basically the CEOs of each of these agencies, they have massive unilateral control and, apparently, can simply choose to unmake the agency at any point.
Except unlike CEOs, they require regular voter approval or they get punted out. If they make a stink the public hates the next president can easily axe their ass. That alone is infinitely better; how many of these horrible companies have 10-20 year CEOs pushing all this shit?
>drag their CEO into congress.
So they can lie and stick their tongue out for days, or just not show up? There is never any consequence for this, congress is functionally toothless. Only the courts and the executive can actually use punitive action.
The fact that there is an avenue of redress at all for public officials makes them so much more accountable than any corporation. Which is exactly why corporations spend so much money influencing public election and congress, they need to subvert the only system than can keep them in check.
Want to create new definitions for new words or phrases? Totally fine, but then Sony would need to replace the “buy” button with “blease” or “blent” or whatever they want to call this fake purchasing shit that all software has been playing at for entirely too long.
Or just be honest and call it leasing.