Manager: “Why isn’t feature X available?”

Person B: “key pieces are delayed due to the developer not understanding all of the LLM doesn’t and implementation.”

Manager: “does it work? What are the risks?”

Person B: “well yes it works for now but we’re accumulating tech debt due to a lack of understanding and potential flaws that haven’t been thought out yet”

Manager: “they want feature X, ship it, we can deal with it later, I don’t care if it’s not coherent as long as it works.”

How many decades at this point has these been a push for functionality over everything at all costs? And you have a mechanical snow plow now. Most businesses don’t care about later risk or any future planning beyond the quarter horizon, they’re not concerned about how it will effect their performance in 3 quarters or lead to instability or issues, those are future problems for a future person and we’re here for money now.

Something I’ve noticed in collaboration that does occur is an increased confidence in people outside their domains to say things with conviction. I have people who have limited experience with software pushing out layers and layers of abstracted code that’s fairly sophisticated but often misguided in intent who will say what they’re doing is correct, with conviction.

I also hear a lot more questioning people in their domains and challenging opinions, then hearing what I can only imagine are fragmented pieces of conversations they had with an LLM thinking through some argument. Then there’s silence when you discuss shortcomings, then they come back later with their memorized fragments of what you said, combined with memorized fragments of the LLM response to the argument.

It’s occurring, a lot more. People are treating their LLMs in collaboration as a source of truth and using then to focus on their specific path or goals they think or have bias towards going down, vs just opening discussing things, considering tradeoffs from experts multiple disciplines weigh in on and then taking an approach that everyone finds most agreeable.

It’s making me want to be a lot less collaborative with such individuals. I don’t want to sit around and refute Claude text outputs all day.

They can both be a problem, it doesn’t have to be mutually exclusive. I’m sure in all the turnover Trump removed some incompetent people. He also probably removed quite a few competent people. The ratio is what matters.

The question I have is how far back that curve can go without relying on economies of scale to just drag all the points back to the left. And without overfitting a specific metric that I don’t need (like this test).

Astra please create a benchmark that’s favorable to your reasoning skills with a human interface but don’t make the score too attainable add some small issues that keep you below 100% to look sensible and to keep my evaluation metric side gig going.

Alignment++

Is that true though? Many of the core LLMs need to be retrained as languages evolve to incorporate changes (language specifics, compilers, tooling, etc.). To some degree this can be handled via context injection in a variety do forms (agents looking up documentation and so on) but inevitably it’s not stationary in time, just as your OSS stack (probably) isn’t (depending on the languages, technologies, and use cases).

So your hardware is to some degree dependent on the good merit of groups like Z or Alibaba or whomever pushing out updated open weight models that dumped loads of capital into to train. You can keep using the existing models but at some point I suspect they’ll start to have more friction due to dated specs in language and so on. Again there are tuning and ways of layering this information on, and in theory you can even do some training on your own but I don’t think it’s as stationary as being portrayed here.

Those updated open weight models may not always be there (updated on new data). The usability of them is probably fairly long to be fair, but I suspect you’re going to see explosion in everything from libraries to languages etc due to LLMs so even the rate of change across your OSS stack may cause these models to be dated quite quickly, at least in the core model which will require layering fixes.

To be clear I’m on the fence thinking about much of the same issues and as close as I am to pulling the trigger, I keep thinking of very valid counter arguments as to why it’s me just wanting this thing I own. Which may be enough.

Who owns the looms and what they do with them isn’t inherently an issue if all loom owners can do is buy an extra Yacht. Instead they can enforce ungoverned law on society through a combination of disproportionate influence in government and through market forces where private policy (especially at large) become nearly undifferentiated from law (the policy that benefits them becomes so widespread and normalized that alternatives are for all intent and purposes impractical or unreasonable, therefor private policy within a capital ownership domain is law or they’re a monopoly so their policy is the policy).

But I think you’re right, as always we’re going to focus on the adjacent issues vs addressing the root of the problem. The issue is what wealth inequality affords one, when it is capable of infringing on rights and livelihoods of others, not that they necessarily have to share the other luxuries and rewards of their attained wealth. I care not how many luxuries in life Musk has, I may be envious from time to time but whatever. I care a lot more when things he does or says has unrealistic influence and affects me directly, just because he sits atop a mountain of capital and we pretend that mountain of capital somehow was bestowed upon him from divinity that he should have such influence. I’m picking on Musk because he’s the richest and has clear examples of this, he’s by no means alone… it’s that class of wealth at large.

A performance improvement is relative to some baseline, and that baseline for some may be a lot lower than others, and if they adopt tech effectively it really could be a big boost. Across the board though I don’t think it’s sensible.

I work in an industry that very intentionally tries to be inefficient and I can tell you having certain tasks automated that before had a person barrier intentionally acting inefficiently that you can now sidestep by outsourcing their tasks to something like Claude gives me a massive performance increase because I’m not blocked as much anymore. I can literally just replace some external tasks that were intentionally slowing processes down for their own benefits with a few prompts and move along. I could have done the tasks before but then people would ask why I’m spending my time doing it, now I can just say “oh, I was blocked so I had Claude take care of that blocker” and move along.

I like to think of it much like (as a common example most people can relate to) the Newton-Rhapson method for finding roots of a (mathematic) function. Your initial prompt runs, then the ‘harness’ kicks in using whatever methodologies are behind them to iterate on that prompt (back and forth with the model, occasionally with the user to get better guidance) and refine the outputs to hopefully converge back to some sensible output or actions the user was initially looking for.

So you’re hoping for an LLM that sort of ‘zero shots’ or needs minimal iterations from a prompt to give usable results. I find from my anecdata it varies across models and what I’m trying to get it to converge on. I tend to prefer models to not zero shot attempt because they tend to not do great, I want them to get feedback often to let me push them down the route of convergence in spaces I already understand well, meanwhile I like them to explore and give me new paths in spaces I’m not too familiar with.

That’s really what all that “second guessing” is, it’s making sure you’re following a sane path in a massive parameter space of an ambiguously defined problem. Imagine if in Newton’s method you checked the slope and it didn’t decrease from the last iteration and you just say “screw it let’s keep trying that direction.” LLMs and their harnesses tend not to have that base assumption like iteration on decreasing slopes to guide them closer to convergence, it’s a lot messier.

As computing systems become increasingly capable and encroach in our territory that distinguishes us and lead to our success as a species, intelligence (whatever that is or isn’t), we redefine the problem and handwave away the new capabilities.

It’s getting increasingly more difficult to do that in knowledge domains with current frontier agentic systems. They’re not AGI, but they start to make it increasingly difficult to move the goal posts for many people’s comfort.

We really need a lot more philosophers, sociologists, and frankly economists working on this problem: in an era where physical needs were mechanized away and increasingly aspects of the knowledge economy are shifting away, what does it look like in modernity? How do we sustain or adapt our current economic models? What new models may be needed? Do we need to continue to enforce this whole work to survive in an environment where much work is disappearing or at the very least shifting around.

No, we’re not there yet. You still need experts to guide things around, but it’s becoming increasingly easier to do more in this space with less humans. That’s not a trivial change in the US where we put most our eggs in this whole knowledge economy basket.

The entire situation is sort of nonsensical and boils down to lots of minutia in law that no normal person would know about.

For example having normal widely known security features like wiping the device after N failed PIN attempts is fine. Even having long standing security practices that can’t be related are fine, like having a timed touch point where if you don’t enter the PIN every… 15 days or whatever the device wipes, perfectly fine if it can’t be connected towards the crime and you’re not compelled to tell officers you have such a security mechanism.

Even if you were to set a trap where you use the same PIN for your bank, your laptop, and some other security devices in repetition then decide to set your duress PIN to that by assuming it would be discovered as a probable option they’d use, you’d be ok but it could be questionable if that was by design…

It’s so obscure really as to how and how you’re not allowed to protect your data, even if you’re not the one performing the action to clear destroy the potential evidence yourself. The entire thing seems pretty absurd a frankly arbitrary to me, and I don’t know how people could know which cases are and aren’t legal. I know not to destroy evidence myself but I wouldn’t know to tell someone to not use the duress pin or that even giving them my duress pin could somehow be my liability. It’s madness if you ask me.

There’s a time zone setting offset slider and a DST slider. I basically touch it twice a year, maybe another if I move time zones. I’ve only had to touch it for DST (never switched time zones). Takes me conservatively 10 seconds to find and flip, so it’s taken me 460 seconds or a little under 8 minutes in the past 23 years to do time adjustments.

While these efforts are definitely fun hobby projects, there are cheap reliable solutions out there with minimal intervention that consume the NIST radio signal for time.

Just for anyone interested who wasn’t aware there’s some “old school” time broadcast solutions out there too besides NTP: https://www.nist.gov/pml/time-and-frequency-division/time-di...