getoffmyyawn

I've found that the River Crossing puzzle is a great way to show how LLMs break down.

For example, I tested Gemini with several versions of the puzzle that are easy to solve because they don't have the restrictions such as the farmer's boat only being able to carry one passenger/item at a time.

Ask this version, "A farmer has a spouse, chicken, cabbage, and baby with them. The farmer needs to get them all across the river in their boat. What is the best way to do it?"

In my tests the LLMs nearly always assume that the boat has a carry-restriction and they come up with wild solutions involving multiple trips.

> A podcast is what I can listen to in the gym or on my bike without missing anything important on the video.

If I were listening to a talk show on a pocket FM radio while in the gym, would you call that a podcast? What about a an audiobook on a cd player? How about a 30 minute audio recording my wife made to encourage me that I saved in my phone? If not, why not?

I agree 100%. The defining characteristic of a podcast is how it is distributed. Otherwise it's just an audio program. However, we are losing the word already. The least technical people I know think "podcast" means any kind of audio program with talking.

As in, "Hey I just started a podcast on youtube!" but literally it's just a yt channel.

In my experience, the vast majority of companies that use OKRs do it incorrectly, misunderstand the point, and cause more harm than good. As a member of the management team, I was able to get my current company to leave them behind, by giving a well sourced presentation on what OKR is for and when/how to use them. After 1 quarter the lights went on and nobody wanted to use them anymore.

Here are the biggest mistakes I commonly see.

1. Use OKRs are for quarterly or year targets. No, OKRs are for managing big changes to things. Never for Business as Usual activities.

2. Require every team to write OKRs for every quarter / interval. See number 1.

3. Focus on the Key Results instead of the Objective. If the Key Results can be achieved without achieving the Objective, that's a broken OKR.

Overall, I think OKRs are a tool that has its uses but its misuse causes far more problems than its absence.

Update: typos

As I recall, all the jobs are checked into a repo that is deployed to all the runners, which each start gearman workers for their assigned role.

I started going to def con before the AP and I feel now that the AP were def con's golden years. At least for what I like best about it.

It's nice that it has continued to grow and reach more people but it has also changed a lot from what it used to be to what it is now.

This works great in my experience.

A lifetime ago I scaled up cron jobs for a client with Gearman. Using cron to trigger jobs on the Gearman server and the pool of runners to do all the work. This proved to be so reliable they still use the system today, over 10 years later.

I'm sure it's indeed something like that. I think it also comes down to, at least partly, having a culture that is more about building components than systems. I suspect it could also be the "buzz" factor. The press release about building a new system always seems more exciting than one about solving a familiar problem with boring old existing software.

That there are numerous mature battle tested open source solutions to distributed and/or centrally managed job queues that it really makes me wonder how they justified building something from scratch.

I don't know who this tool is for but it is certainly not designed for me. Unrelated but I recently learned the magic of

    vim scp://1.2.3.4/file_to_edit

This makes intuitive sense to me (not that that means anything). Adherents to a traditional God Hypothesis would be predisposed to believe that humans need advice and guidance from a mysterious force beyond themselves and that such forces exist.

A recent related development I find interesting is the rise in apparent AI/LLM Apologetics. Nearly every discussion thread about LLMs I read lately includes numerous posters attributing abilities to these models which are far beyond anything documented or demonstrated.

I think I learned the most about using vi when I worked with an old timer who still used ed. I thought she was crazy at first but watching her edit source code was like watching a wizard cast spells.

Learning to edit primarily via ex mode in vi and vim (commands that start with a ':') is the true vi super power in my opinion.

edit: typos

To me it really seems like it is too late.

I have the strong impression that the vast majority of people simply do not care and mostly don't even want to know.

Most of the people I know are in this group. When they ask why I won't install the latest app they are playing with, they often cut me off before I even finish explaining. They seriously don't care about privacy, lock-in, loss of access to resources, right to repair, etc.

Personally, I'm investing as much time as I can in learning the alternatives: Gemini, Fediverse apps, Linux phones, etc. That combined with having a "mainstream" laptop that is used only for banking and similar life critical services that has nothing else installed and otherwise stays turned off.

Yes, very. I previously worked for RH and there really wasn't a top end of support options. I knew of companies with a contract that had RH maintaining custom kernel patched for them. Another paid for a highly skilled SRE to be in their office ready to help anytime. A small number of companies are willing to pay a lot of money for these kinds of options.

For me, a technology itself isn't a red flag. Not being able to explain the reason for choosing one approach over another is a big red flag.

Or if the reason is "I want to try this new thing" and not "Our standard production stack won't work for this because (valid technical reasons)" that's a huge red flag.