How do you actually incorporate these into your prompt? Because I've tried and I've seen many others try and report that it does not work well, especially over long sessions.
I also don't know how much to trust the model, but I've had the model tell me specifically that certain aspects of ASD-STE100 are unactionable and will just create more noise.
The OP's own skill even leads with something in a very similar vein:
> These rules apply to every response for the rest of the session, not only this one. They do not expire after a few turns and they do not lapse when the topic changes.
My understanding is that phrases like this are at best a _very_ weak signal to the model. It's simply contradictory to how the model works at a level that can't be overridden by injecting tokens.
Weak is even giving it too much credit imho, a yapping skill telling Claude to not yap can’t but fail.
I placed those instructions in setting > personal instructions, in my global CLAUDE.md, and in each project’s CLAUDE.md. I also use the concise output style in CC. If I choose a high thinking model it will start deviating in long sessions, then I just remind it in my next message:
“Remember ASD-STE100 style.
[rest of my message]”
BLUF sticks a lot easier than STE to be honest… but Claude knows what STE is, and using the “ASD-STE100 style” locution avoids the compliance issue (it’s true that strict ASD-STE100 compliance isn’t really possible, nor desired)
I actually wish the model would go completely in the opposite direction. Except in the rare circumstances where the model has actually measured something, it is hilariously deficient in its concept of time. It will often output phrases like "this relates to <thing> that you did weeks ago", referring to something that happened in the session just a few turns (and hours or a couple of days) ago. Likewise for estimating how long a coding tasks takes, it is hilariously inept. It honestly feels like it rolls two completely independent dice to select a number and a value from (hours | days | weeks) when it needs to attach an estimate to something. I'd much rather read "a bit" than be distracted by these utterly nonsensical times.
The full skill gives this example:
>Bad: "This will take some work." Good: "About 15 minutes if tests already cover this. An afternoon if not."
My experience is that it is very likely that whatever task this is describing takes anywhere from 1 to 30 minutes, consistently. Maybe I just work way faster than the average person.
I think I'd be a strong fit for some of your roles, but the benefits advertised read to me like the absolute bare minimum to meet industry norms. Is there anything you would say is a real advantage for Apex?
I really hate that I'm even typing these words right now, but wouldn't this be one of the few fields where modern LLMs would potentially provide a real capability that was previously lacking? I'm imagining an LLM might be able to provide rough first pass translations and provide hints about which specific tablets might be most interesting for a human that can't possibly look at them all.
"Artificial intelligence can therefore provide researchers with a significant support, helping them link together texts that mention the same individual, a geographical site, or a word within a large digitised corpus. But we are still far from machine decipherment of texts."
I doubt there's much Akkadian and the 14 other languages written in cuneiform over more than 2,000 years. Several of these are language isolates. Its probably not intractable but would require a lot of human labeling and correcting by academics who can read these languages. It would be a fun legacy project for an eccentric billionaire. If I were Bezos, it would be on my todo list.
and imo ive also thought the same thing, but then it does still come down to the amount of people and time it would take to go over those rough translations and pull the wanted info out of it. manpower is lacking in a lot of those fields
Why would you hate that you're typing those words? Technology is useful and neutral, and if we have good uses for it like it seems so here, let's do it! :)
A decade ago every "omg just use [answer]!" was to suggest blockchain solve your problem.
Every possible problem seemed to have a solution that involved blockchain.
... now that "[answer]" is always "AI".
So I get his sense of cringe... while also acknowledging I had the same thought. We've never been better at pattern matching, pattern identification, OCR, translating across languages... seems like a great use of an LLM.
The United States population has more than doubled since 1950. There is no "before" that we could actually go back to. The massive, sustained population growth is built on the massive efficiency growth.
Everyone has different opinions about this stuff, but I for one found _working_ in the office to absolutely be the worst part. And since Covid, my workplace has torn down old buildings and reduced square footage per employee, so if I went back to the office it would be even worse.
It’s been a broader trend of businesses downsizing the square footage when they can and when they can’t they start to mandate in office to justify the cost.
I don't know how broad the trend is, but I am a contractor for the US government. I'm told that agencies are actually required by law to reduce square footage. So when they tore down my old building (which absolutely needed to go), they simply could not as a matter of law build a new building with the same square footage even if they had the budget.
