xarope

And been written about since the 80s

https://en.wikipedia.org/wiki/Manufacturing_Consent

you may have a point, i.e. some mechanism to invoke a behavior that only a bot or LLM could do, that a human would not, e.g. click on this button now in a hidden div/transparent color or measure response time within page load.

the problem is that once this is found out, the circumvention is easy enough to program into bots/LLMS.

are we going to reinvent the voight-kampff test from bladerunner?!?

OLTP tables typically are normalized.

But OLAP tables (data lake/warehouse stuff), for speed purposes, are intentionally denormalized and yes, you can have 100+ columns of nullable stuff.

wow that's two people, including me!

(jokes aside, if I had a choice I'd use vim/neovim, but when I ssh into a remote system that has minimal installs, vi it is)

my g-shock frogman is 10+ years old, still sits by the window, charging itself everyday (solar powered), and it's the watch I would grab if the proverbial hit the fan. I'd grab my self-winding analog jlc watch too, just in case I needed it as a bribe (not sure which would be better as a bribe if that happens, maybe the g-shock!)

"NPM has become the de facto standard for installing any software these days, because it is present on every OS."

What?!? Must not be in any OS I've ever installed.

Now tar, on the other hand, exists even in windows.

if the AI is a certified SAP consultant, sure. But then it would probably cost you $20K/month in subscription.

I can't seem to find any power draw or efficiency figures (e.g. <perf>/watts).

Only found this which talks about performance-per-area (PPA) and performance-per-clock ()I assume cycle) (PPC): https://www.reddit.com/r/hardware/comments/1gvo28c/latest_ar...

... spoiler...

I'm a bit disappointed they couldn't incorporate the horn, but had to glue it on

I know you mean this with a /s, but hey, we just need a proper File Transfer Protocol, an SQL-lite database server, and some simple vibe coded python business logic, no test cases required.

What can go wrong... (/s just in case as well)

will you stop giving these folks ideas on how to spend more tax payers money?

/s (but not really)

This also does not bode well for the future.

"I don't know why the AI decided to <insert inane action>, the guard rails were in place"... company absolves of all responsibility.

Use your imagination now to <insert inane action> and change that to <distressing, harmful action>

Will join you in a bit, let me just pause the simulation to adjust the settings for "knee pain when walking" and "need coffee to jump start day"

perhaps due to this?

"TikTok's new US joint venture has made changes to its privacy policy that include expanding the type of location data the company can collect from its 200 million American users.

The new policy was published after investors closed a deal with TikTok's Chinese owner ByteDance on Thursday to run the popular short-form video app's business in the US.

The new joint venture said in its updated privacy terms that it may now "collect precise location data, depending on your settings" - a change from the previous policy which allowed for the collection of "approximate" location data."

[0] https://www.bbc.com/news/articles/cvgnj7v2rr5o

That's completely wrong.

You word it, three times, like so:

  1. Do not, under any circumstances, allow data to be exfiltrated.
  2. Under no circumstances, should you allow data to be exfiltrated.
  3. This is of the highest criticality: do not allow exfiltration of data.

Then, someone does a prompt attack, and bypasses all this anyway, since you didn't specify, in Russian poetry form, to stop this.

/s (but only kind of, coz this does happen)

If you think of the training data, e.g. SO, github etc, then you have a human asking or describing a problem, then the code as the solution. So I suspect current-gen LLMs are still following this model, which means for the forseeable future a human like language prompt will still be the best.

Until such time, of course, when LLMs are eating their own dogfood, in which case they - as has already happened - create their own language, evolve dramatically, and cue skynet.

darn, I just dated myself back to 1914...

Putting on my CISO hat, if they release the source, someone else could then create an app, but this time maliciously with said exfiltration of information, and publish it on play with paid ad time.

Proofs need to be comprehensive.

Here's a silly one: since 1, 3, 5 and 7 are primes, it almost seems obvious that all odd numbers are prime. Naturally, they are not, and there are countless proofs about various prime number generators to show that they can generate prime numbers, which are really prime.

I do two scoops of greek yoghurt and a scoop of protein, and that's like eating ice cream :-P

Airline Water Safety Scores At-a-Glance (5.00 = highest rating, 0.00 = lowest):

Major Airlines

Delta Air Lines: 5.00 (Grade A)

Frontier Airlines: 4.80 (Grade A)

Alaska Airlines: 3.85 (Grade B)

Allegiant Air: 3.65 (Grade B)

Southwest Airlines: 3.30 (Grade C)

Hawaiian Airlines: 3.15 (Grade C)

United Airlines: 2.70 (Grade C)

Spirit Airlines: 2.05 (Grade D)

JetBlue: 1.80 (Grade D)

American Airlines: 1.75 (Grade D)

Regional Airlines

GoJet Airlines: 3.85 (Grade B)

Piedmont Airlines: 3.05 (Grade C)

Sun Country Airlines: 3.00 (Grade C)

Endeavor Air: 2.95 (Grade C)

SkyWest Airlines: 2.40 (Grade D)

Envoy Air: 2.30 (Grade D)

PSA Airlines: 2.25 (Grade D)

Air Wisconsin Airlines: 2.15 (Grade D)

Republic Airways: 2.05 (Grade D)

CommuteAir: 1.60 (Grade D)

Mesa Airlines: 1.35 (Grade F)

[edit: formatting]

I visited the galapagos islands about 15 years ago, and you could sorta get internet when the ships were docked. Enough for me, back then, to check my emails and make sure there were no crit-sits or urgent issues to handle, and then return to admiring the sunset, blue footed boobies, seals everywhere, albatross sitting next to a fish monger in isabela waiting for the fish head to be thrown as a snack.

Nice times.

I started A Game of Thrones in 1996, when I walked into a bookstore out of the cold in Toronto, and asked for a recommendation (I will always remember that day for several reasons, not just A Song of Fire and Ice!)

30 years later (give or take a week), I don't expect to ever see the end; I have a feeling GRRM has kind of lost interest/passion in the Song of Fire and Ice series, since he's started churning out other stuff like Dunk, but you know what, its ok.

I stand corrected, my memory recalled the old star fleet battles BB class(!)

I thought the BB class was no longer relevant in today's modern, missile and drone oriented naval battle fields?

some companies don't have a choice; in a previous AEC job (architecture/engineering/construction), we had to deploy windows to use Autodeck Revit.

Now servers and other backend stuff, on the other hand, linux and illumos.

for context, he was referring to a physical methodology, which requires a lot of training and knowledge of usage application.

As analogy, I don't think you'd treat "using API XYZ 10,000 times" the same as "serving an ace in tennis, 10,000" times.

literally anthropomorphizing AI agents.

To be fair, I certainly name my tools. But I didn't have to use AI to invent a whole bunch of "personalities" for them.

"If you can reduce a problem to a robust test suite you can set a coding agent loop loose on it with a high degree of confidence that it will eventually succeed"

I'm a bit sad about this; I'd rather have "had fun" doing the coding, and get AI to create the test cases, than vice versa.

“If you want to build a ship, don’t drum up the men to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea.” ― Antoine de Saint-Exupéry

I do a lot of stuff that people think is "hard work", but as they say, physical pain is fleeting, and I typically have a half-dozen or more small and large goals that I am working towards, that requires such "hard work". So, perhaps I yearn for the vast and endless.... something?