datsci_est_2015

> To put humans through this kind of misaligned system is the stuff of nightmares.

And also the natural market equilibrium, apparently.

> Yes, the average distance per run may be higher for the 5 companies, but it won't be 5x.

You contradict yourself at the end of the comment. Yes, it could be estimated a linear equation (mx + b) where m is the cost per house and b is the cost per route, roughly. And then you could have a system of which equations, where each row in the system’s matrix corresponds to an additional company.

The optimal result is probably not 1 company, but it’s probably not 1 company per house either

Poor fella doesn’t have a seat. Intriguing design where both pedals are on the same side of the frame. Balancing must be a challenge.

Network administrators also solved these things, e.g. you can limit what websites a virtual machine is allowed to access.

Now, if you’re letting your agents abuse network traffic there becomes a point where you’re criminally liable. Unfortunately the current US government has no interest in prosecuting such abuse.

I can’t recreate a shuffling of cards.

Combinatorics is basically the study of mathematically generating and quantifying complexity.

To be fair, after visiting some trade shows recently, a lot of these “parasitic” LLM startups are not very convincing from a business-use case perspective. (They’re not really parasitic, more like a remora attaching itself to a shark, the shark being OpenAI or Anthropic).

At IMTS at Chicago the skepticism amongst the visitors towards AI and pure-AI startups was at an all-time high. More than once I heard, “Oh yeah I visited that AI booth and they couldn’t explain what value they would add to my business.”

So in that sense, while MCPs are “winning”, I also see them as a significant part of the AI “bubble”, specifically solutions looking for problems backed by VC money.

Why trust any user with access to any terminal? UNIX solved this decades ago. It’s just we’re all learning to be sysadmins for the most creatively destructive and persistent set of users of all time.

Wow, that quote goes hard.

I’m reminded that almost no one beyond a select few knew high up in the military and around the emperor knew how badly the Japanese were defeated at Midway.

Paternalistic. Arrogant. Shameful. And deeply engrained in the Japanese cultural zeitgeist (of the early-mid 20th century).

Edit: I guess it’s commonly attributed to a German citizen, but their cultures mirrored each other. Fascism falling under the weight of its own propaganda.

Maybe the implication is that Warren thinks it’s significantly easier to run a wealth fund than get gold in the Olympics. I would probably agree.

Well the current leaders of the US Government have demonstrated that they believe that people of certain races, sexual orientations, and genders are less than human - and enthusiastically engage with other leaders who have similar beliefs. It’s because of this, along with their arrogance and incompetence, that I am personally not the least bit surprised that they have committed and will continue to commit war crimes.

I guess the trick for me is (1) convincing others and (2) convincing others to give a shit.

We could enforce (suggest?) a common format / api at the federal level. Especially if it’s incentivized with funding that more than justifies the cost of maintenance. Similar to how federal interstate funding is only available to states with a 21+ drinking age.

You don’t even have to limit it to machine learning, the definition of forecasting is isomorphic to the definition of modeling, which, with the dilution of the term AI, is also isomorphic to the definition of AI.

More simply:

  - forecasting = modeling = AI

Edit: I’d even throw statistics into that extended equality, meaning that Bayes, Bernoulli and even the fellow named John Gaunt have a strong case for having invented AI.

Bring this energetic criticism onto Warsh, whose speech was nauseatingly political in comparison to his predecessor. Paraphrasing ~ “I’m not data-driven, I’m trend-driven”. Frankly speaking, he sounded like a dumbass who failed his way upwards into a position of meaningful power, much like the rest of the government at the moment. Warsh also refused to give a target or a forecast.

The OP that you’re criticizing is simply making a prediction, and he doesn’t work at the fed.

Yes all of my the customers of my accounting software love the ambiguity of how it will react to them processing entries.

You’re not saying anything with substance, but I guess that’s not surprising given your stance on LLMs.

A 6 year old will figure out how to ask you to help them after they get it wrong.

Maybe watch some HuskIRL videos to temper your expectations. Sure, frontier models providers may alter their harnesses to better target chess, but that’s lipstick on a pig imo. The models themselves are not, in isolation, capable of solving general tasks. We haven’t modeled intelligence sufficiently. We’re in a local minimum and throwing billions of dollars at a gamble that that local minimum can facilitate the concentration of wealth even further and fully realize the American dream of eliminating the middle class.

Why does a 6 year old not need any of these guardrails?

Frontier model’s failure modes are a direct refutation of claims that we’ve reached (or will soon reach) the artificial general intelligence. We may have reached an artificial general intelligence, but there may be more complexity to this than even AI thought leaders are talking / influencing about.

Maybe not all AGIs have a path to digital singularity. Maybe our current era of intelligence modeling has fundamental flaws and we are in a local minimum of the artificial intelligence space.

To note, I would bet with a good amount of certainty that we have enough compute power and automation to DDOS the internet out of existence with botnets. That doesn’t make the frontier models intelligent, that just makes their handlers reckless.

Funniest comment I’ve read all day, thanks. Never seen someone make light of a government rendering someone stateless. Especially rich in the context of an apartheid state.

And yet they still hold immense power to make our lives shittier. It is far easier to destroy a functioning system than it is to build one from the ground up

The last 10% is probably the most important. It sounds like you have agents doing all the tasks that SaaS would’ve provided in the past. Most people wouldn’t have shelled out for that SaaS though because the subscriptions would eat their margin. You probably still (occasionally) need things like legal counsel that would be foolish to use LLM output for.

If you have an agent that can run a business, why would I pay you instead of running my own agent that provides the service your business provides?

The only niche where that doesn’t work is capital investments. I can’t have an agent that provides me housing, but your agent can run your slumlord business.

Maybe very passive businesses which are basically just capital investments with some paper pushing (e.g. real estate slumlord).

But these types of businesses don’t really create value as much as seek rent.

Tell that to the script kiddies with a criminal record for "hacking" into their school's computer systems by entering "username: admin" and "password: password".

> Your idea that datacenters are racists transitively is woke slop.

Your knee jerk reaction to refer to something as “woke slop” belies your privileged distance from anything remotely resembling hardship (I LiVe iN LoUdOuN) and an underdeveloped sense of empathy.

Anyway, we should be extremely skeptical of how infrastructure is built out in the US because of historical baggage. Poverty and industrial development create a vicious cycle that leads people to walk around with their thumb in their ass and claim that the poors just aren’t pulling themselves up by their bootstraps hard enough.

Code quality is probably isomorphic to the halting problem, or can be reduced to the halting problem in the simplest case. I.e. it’s intractable.

So how will separate systems communicate with each other? Or are there no separate systems? That’s the digital singularity I suppose.

This thread started by referencing grad students, and I can vouch for it from personal experience with physics grad students.

This is a very insightful dynamic.

Probably reinforces that we’ve already surpassed the frontier threshold for LLM usability in software development and can now focus on cost and personalization. To make a comparison, no one is making a better machine vision app for hot dog classification - we hit diminishing returns 10 years ago on that front.

But also scary for both investors and the working class: AI companies want to facilitate the concentration of capital even further into the hands of the ownership class. Will they succeed?

+1

Some people write terrible code. Some people don’t proofread their own code. Some people are writing code in a second (or third!) language and the typos are harder to spot and comments are harder to formulate.

It’s a bit like the variance you could expect from asking a room full of people to write a 1-page short story on a specific topic (e.g. “death of a loved one”). Some of those short stories will be unreadable, and some of the people who wrote those terrible stories will have no idea how bad they are. Except now, enter AI, and the room generates the short story instead - who has the skills to determine whether their short story isn’t terrible?

Maybe AI will provide an impetus for change. One can only hope.

I’d be willing to pay a few hundred dollars a year to be a part of a professional organization that maintains ethics requirements at this point.