641–650 of 706 posts
Earlier quoted context omitted.
What kind of security issues are you thinking about? I'm generating UI components like Selects for certain data types or Charts of data.
Do these components have JS, do they have npm dependencies? Since AI slopsquatting is a thing https://en.wikipedia.org/wiki/Slopsquatting
This is definitely one of my CORE problem as I use these tools for "professional software engineering." I really desperately need LLMs to maintain extremely effective context and it's not actually that interesting to see a new model that's marginally better than the next one (for my day-to-day). However. Price is king. Allowing me to flood the context window with my code base is great, but given that the price has su…
Anthropic also recently said that they think that longer/compressed context can serve as an alternative (not sure what was the exact wording/characterization they used) to continual/incremental learning, so context space is also going to be competing with model interaction history if you want to avoid groundhog day and continually having to tell/correct the model the same things over and over.
It seems we're now firmly in the productization phase of LLM development, as opposed to seeing much fundamental improvement (other than math olympiad etc "benchmark" results, released to give the impression of progress). Yannic Kilcher is right, "AGI is not coming", at least not in the form of an enhanced LLM. Demis Hassabis' very recent estimate was for 50% chance of AGI by 2030 (i.e. still 15 years out).
While we're waiting for AGI, it seems a better approach to needing everything in context would be to lean more heavily on tool use, perhaps more similar to how a human works - we don't memorize the entire code base (at least not in terms of complete line-by-line detail, even though we may have a pretty clear overview of a 10K LOC codebase while we're in the middle of development) but rather rely on tools like grep and ctags to locate relevant parts of source code on an as-needed basis.
Earlier quoted context omitted.
is Opus that much better than Sonnet? My sub is $20 a month, so I guess I'd have to buy that I'm going to get a 10x boost, which seems dubious
Yes, Opus is much better at complicated architecture
Earlier quoted context omitted.
My ability to break a problem down does not start from listing the files out and reading a few. If you're completely new to the problem then ... yes, it does. You're assuming that you're working on a project that you've spent time on and learned the domain for, and then you're comparing that to an LLM being prompted to look at a codebase with the context of the files. Those things are not the same though. A closer an…
> If you're completely new to the problem then ... yes, it does. Of course, because I am not new to the problem, whereas an LLM is new to it every new prompt. I am not really trying to find a fair comparison because I believe humans have an unfair advantage in this instance, and am trying to make that point, rather than compare like for like abilities. I think we'll find even with all the context clues from MCPs and…
That is true for the LLMs you have access to now. Now imagine if the LLM had been trained on your entire code base. And not just the code, but the entire commit history, commit messages and also all of your external design docs. And code and docs from all relevant projects. That LLM would not be new to the problem every prompt. Basically, imagine that you fine-tuned an LLM for your specific project. You will eventually have access to such an LLM.
Earlier quoted context omitted.
https://research.trychroma.com/context-rot
This is a good piece. Clearly it's a pretty complex problem and the intuitive result a layman engineer like myself might expect doesn't reflect the reality of LLMs. Regex works as reliably on 20 characters as it does 2m characters; the only difference is speed. I've learned this will probably _never_ be the case with LLMs, there will forever exist some level of epistemic doubt in its result. When they announced Big C…
As the piece above references, this is a totally insufficient test for the real world. Things like "find two unrelated facts tied together by a question, then perform reasoning based on them" are much harder.
Scaling context properly is O(n^2). I'm not really up to date on what people are doing to combat this, but I find it hard to believe the jump from 100k -> 1m context window involved a 100x (10^2) slowdown, so they're probably taking some shortcut.
Earlier quoted context omitted.
Strong agree. Bash is so annoying that there have been many scripts that I wanted to have, but just didn't write (did the thing manually instead) rather than go down the rabbit hole of Bash nonsense. LLMs turn this on its head. I probably have LLMs write 1-2 bash scripts a week now, that I commit to git for use now and later.
Why not use a more sensible shell, e.g. Fish?
Bash scripts are p much universal though. I can send em to my coworkers. I can use them in my awful prod-debugging-helm environment.
This is definitely one of my CORE problem as I use these tools for "professional software engineering." I really desperately need LLMs to maintain extremely effective context and it's not actually that interesting to see a new model that's marginally better than the next one (for my day-to-day). However. Price is king. Allowing me to flood the context window with my code base is great, but given that the price has su…
Even 1 MB context is only roughly 20K LOC so pretty limiting, especially if you're also trying to fit API documents or any other lengthy material into the context. Anthropic also recently said that they think that longer/compressed context can serve as an alternative (not sure what was the exact wording/characterization they used) to continual/incremental learning, so context space is also going to be competing with…
2030 is only 5 years out
Earlier quoted context omitted.
I have an entire life worth of context and I still remember projects I worked on 15 years ago.
Not with pixel perfect accuracy. You vaguely remember, although it may not feel like that because your brain fills in the details (hallucinates) as you recall. The comparisons are closer than you might think.
Also, you're a programmer you have no foundation of knowledge on which to make that assessment. You might as well opine on quarks or martian cellular life. My god the arrogance of people in my industry.
This is definitely one of my CORE problem as I use these tools for "professional software engineering." I really desperately need LLMs to maintain extremely effective context and it's not actually that interesting to see a new model that's marginally better than the next one (for my day-to-day). However. Price is king. Allowing me to flood the context window with my code base is great, but given that the price has su…
Even 1 MB context is only roughly 20K LOC so pretty limiting, especially if you're also trying to fit API documents or any other lengthy material into the context. Anthropic also recently said that they think that longer/compressed context can serve as an alternative (not sure what was the exact wording/characterization they used) to continual/incremental learning, so context space is also going to be competing with…
In your working mental model, you have broad understandings of the broader domain. You have broad understandings of the architecture. You summarize broad sections of the program into simpler ideas. module_a does x, module_b does y, insane file c does z, and so on. Then there is the part of the software you're actively working on, where you need more concrete context.
So as you move towards the central task, the context becomes more specific. But the vague outer context is still crucial to the task at hand. Now, you can certainly find ways to summarize this mental model in an input to an LLM, especially with increasing context windows. But we probably need to understand how we would better present these sorts of things to achieve performance similar to a human brain, because the mechanism is very different.
Earlier quoted context omitted.
Even 1 MB context is only roughly 20K LOC so pretty limiting, especially if you're also trying to fit API documents or any other lengthy material into the context. Anthropic also recently said that they think that longer/compressed context can serve as an alternative (not sure what was the exact wording/characterization they used) to continual/incremental learning, so context space is also going to be competing with…
>"Demis Hassabis' very recent estimate was for 50% chance of AGI by 2030 (i.e. still 15 years out)." 2030 is only 5 years out