It can be hard to find the thing you're looking for in documentation, especially if you're not too familiar with the piece of software you are currently looking at. Saving your coworkers a couple of hours by spending ten minutes of your time can be a good trade - even though it's annoying for you.
As I've said, the docs work for many people. Even in a situation where something is hard for someone, I disagree. It reinforces learned helplessness. Yes, work sometimes will be hard, but things get easier as you gain experience. To be fair to myself, the company cares if I achieve my goals and, importantly, does not give me extra time or extra credit to achieve other people's goals. Even with this in mind, as I've already said, I understand the importance of going the extra mile in some cases (to my own detriment). If another employee does not have the initiative or wherewithal to go through the normal channels and is only interested in skipping the line, this is not a good trade for me or our company.
This has nothing to do with LLM's. An 'oracle' is a hypothetical construct that is used in proofs. As the article outlines, useful in complexity theory.
I think you misunderstood GP's very concise point. Allow me to expand.
A developer is a Turing Machine that produces low quality code -- and there is a hierarchy of developers (with the mythical 10X on top) that produce a time hierarchy of time classes depending on how fast they are. Depending on the complexity class you're in, the TM may not be able to produce for example a regex, they might be able to do it in exponential time, and the 10X is able to do it in linear time (linear in the number of characters and bugs). LLMs act as oracle that can produce a regex by just asking (an O(1) operation) so that changes the whole time class hierarchy of developers.
So it's the same as in complexity theory, introducing LLM oracles creates analog of the existing complexity hierarchies.
This is of particular interest for investors looking to reduce developer cost.
Watching GP and parent discuss oracles.. and missing an actual oracle is so HN.. we use imperfect LLM oracles as sounding boards while we theorize about perfect ones.
Oh, the age we live in - of this new math.
LLMs are math that isn't exact, except when it is more exact, and not always when you need it to be. LLM math can't be too accurate or warm or the math doesn't work as well. Like an O(1) operation that randomly decides to run in O(n!).
Humans, technically are just higher maintenance Turing Machines who incrementally write less buggy code.
LLMs get trained on our obi wan buggy odysseys to help them spit out mindbending new ways of regex-ing something and leap to solutions, and chase shiny new things, however it originated.
This shows us we must not only communicate with LLM oracles and perfect Oracles using natural language programming, but to also use the force, just maybe don't trust it to parse HTML. Maybe it knows why it's so hard with to not be able to "ace" any SWE interview to land a high paying job and keep it for more than 12 months.
A developer that ticks all these boxes is certainly a Good Software Engineer, but the reverse relation doesn't necessarily hold. There are many that have made very valuable contributions while not even working on a team, or perhaps even being an asshole to everyone around them, ignoring stakeholders, everything. Or just something less extreme, such as maybe they didn't at all times know their organization that well. That is fine, if it works at their time and place. To call those "Bad Software Engineers" is unhelpful.
I mean, some are claiming that AI will be capable of anything. But I don't think extracting Excel formula's is currently a focus of LLM applications? Do you know of startups or other attempts at exploring this?
Power Apps are pushing into AI direction. And it does use AI to parse excel file. Moreover Power Apps on itself has PowerFx engine that uses Excel formulas for app + more.
Yes I recognize this problem in my organization as well. I think this is feasible, and tools in this direction exist already, like https://formulas.readthedocs.io/en/stable/doc.html. I think one challenge is that the variable names in Excel (B3, B2-10 for a list) are not easily converted to descriptive names.
>I think one challenge is that the variable names in Excel (B3, B2-10 for a list) are not easily converted to descriptive names.
It takes some getting used to, but you can pretty easily create a named range for an individual cell by modifying the value immediately left of the formula bar. You can also setup a table to hold data (insert -> table).
Tables can be renamed and allow formulas like =sum(tbl_salaries[salary]).
With named ranges, your formula can look like =purchase_price*sales_tax
Thank you! I was hoping for a reply like this. The naming is actually (in big corp) a non-problem. I even think many finance workers dislike these models as well, so this isn’t even a bad chore. The models we run have tops a few hundred inputs, many ranged. That’s an hour or two of puzzling.
* Sisu (Finnish): Though not a direct equivalent, "sisu" refers to a blend of determination, resilience, and courage in the face of adversity. It's doing something against the odds, putting extra effort, and not giving up.
* Gaman (Japanese): A term that loosely relates to enduring the seemingly unbearable with patience and dignity. It can apply to doing meticulous, quality work even when situations are challenging.
* Jugaad (Hindi): Jugaad speaks to a creative or innovative fix; essentially finding a low-cost solution to a problem in an intelligent way. It reflects a spirit of resourceful improvisation and can indicate a pride or savvy in being able to solve problems with limited resources.
* Arbejdsglæde (Danish): This word directly translates to "work happiness" and denotes finding joy and satisfaction in the work you do.
* Mānawa (Maori): This is used to describe patience and perseverance, particularly in working toward a goal or mastering a skill.
It depends on the nature of your compute. If it is dominated by IO, or if you are actually calling native libraries (like `numpy` does, or it is something that is handled by `arrow`), there is no reason to switch away from Python. If you are writing custom algorithms, I think https://julialang.org/ is a great option.
