ilovetux

I have yet to see a CEO actually take responsibility for anything. Its always some hand-wavey talk about strategy and unexpected decline in sales. Never once was it "I messed up."

I have, on the other hand, seen boards pile all the blame onto the C-Suite. Seems easy enough to pile the blame on an AI.

True, I tried to convey that.

Thanks. That actually makes it make more sense, chevron doctrine applied to courts and what they can/can't do.

Edit: made applies past tense

Didn't the supreme court recently overturn the Chevron Doctrine?

Doesn't that mean agencies like this are not free to set their own rules when a law is ambiguous?

And isnt the cap on number of households reached actually codified unambiguously in federal law?

Seems to me like this should not be allowed to happen.

Let me know if I am wrong.

First, I would like to preface that this is alpha and should not replace anything in prod at the moment.

What I like about how SIEMatic is built:

* Django (personal familiarity)

* Agent, Indexer, crawlers and web all share a codebase with separate settings

* Search language is built around django orm, built-in jinja2 and ast.literal_eval makes it very versatile, I will be writing some tutorials soon

* search commands designed to handle dataframes, Django Querysets and records (lists of dicts) with predictable conversion between them (you can filter early in you search pipeline)

* from local on sqlite (rundev command) to distributed on postgres with scaling/tuning tips

* comes out of the box ready to monitor your host and provide dashboards and savedsearches

* crawlers handle data retention, alerting, summary event generation and more

There's a bunch more, but again, this is alpha software.

Author here: Thanks for the upvotes!

I built this so I could have an analytics, observability and security stack in my back pocket.

Individuals, nonprofit and educational users are licensed for production use.

Commercial institutions must obtain a commercial license before production use.

Have any questions, feel free to ask.

I struggle with this. A kind of premature optimization. I get around it by taking really messy handwritten notes in a notebook (I guess this is called a dirt notebook, but I just call it my dirt notebook).

I get around this by making a deliberate attempt to organize my notes after I take them. I will usually try to turn them into a word doc with a lot of word art, but if they need to be in a repo or something, I will go with markdown.

Fair enough, I suppose.

While Tesla has (in some years) avoided traditional marketing, the ceo is known for spending ridiculous amounts of money on publicity stunts like having a submarine shipped to a cave and buying Twitter to boost public perception of his companies.

I think this is the exception that proves the rule.

I would agree with you if we were just talking about the abstract idea of marketing. But like everything else, the devil is in the details.

The marketing industry in the US is built not only to get the word out about your product, but also to gatekeep who can compete in our free market.

With marketing being so pervasive as to monetize the entire internet it effectively levies a tax on every business that wants to compete.

If you dont have the marketing budget to outspend your competition then they have no competition.

> "Anthropomorphic" means "human shaped".

In a literal, ancient Greek sense for sure, but in modern English Anthropomorphic would describe the act of attributing human characteristics to non-human entities.

Seems pretty apt for a company that produces one of the more anthropomorphized technologies.

I like the approach outlined in the article. These days having a roadmap for yourself while cruising at highway speeds helps make sense of the chaos.

One big pain point that has existed forever and has never really been addresses adequately is the ability to come up with requirements.

Sure, it sounds easy, I need the app to do x, y and z. But requirements change in real time because of lack of foresight, change of business needs, an unexpected roadblock and more contribute to changing requirements.

So, the advice to come up with the requirements by yourself or with the LLM miss the biggest pain point.

I'd like to see a resurgence of flow charts, IPO (Input, Processing and Output) charts and other tools to organize requirements spring up to help with really nailing down requirements.

I will say, though, some of the pain is relieved because the agent can perform a huge refactor in a couple of minutes, but that opens a whole new can of worms.

I think some AI companies are just now starting to feel the pressure to profit.

Soon, I predict we will see a pretty significant jump in price that will make a 10% productivity gain seem tiny compared to the associated bills.

For now, these companies are trying to reach critical mass so their users are so dependant on their tech that they have to keep paying at least in the short term.

I would push back on that because I have some experience writing marketing copy. It's just not my primary competency.

If the proverbial marketer that you were referring to had some experience with coding, I dont see why they wouldnt be able to review the output and see any obvious flaws.

My whole point is that LLMs are of limited use when you are already an expert or when you know nothing about the subject. However, they really seem to help elevate beginner/intermediate level tangential skillsets.

Obviously everything is still evolving and your results may vary.

It's my opinion that AI can help a lot when it is supporting you when expanding beyond your core competency.

For instance, as a SWE, I get just a little help with boilerplate from the AI. I could usually have done it better, but sometimes the ask is both simple enough and boring enough that the code from the LLM actually produces something very close to what I would produce.

On the other side of the coin, a non-technical person using AI would be unable to properly understand and review the output.

Where it shines is on things that I am OK at. Like writing marketing copy. I can get by myself, but its slightly outside of my wheelhouse, but as long as I have a solid understanding of the product I can use AI to compliment my beginner/intermediate skills and produce something better than I would produce on my own.

A similar thing is writing tutorials. I write some code and documentation, but the tutorials are enough of a slog that I get distracted by my distaste for it. This is a good fit for AI.

I think this is where we will see AI help the most. Where someone's skillset includes the task at hand but at a secondary level where the user might doubt themselves or get distracted with the misery the task brings them.

I've always heard "slow is steady and steady is fast."

Same thing, but from the trades instead of the military.

The insurance company is not a charity. You will (or already have) pay the full price.

> FYI for anyone who isn't familiar with the wacky US insurance situation: Nobody in the US actually pays $800 for the drug. That's the "list price" for insurance companies to pay. Even insurance companies don't pay that price because they negotiate their own rates with the drug companies, which are lower.

Sure, we do not pay $800 at the pharmacy when we go to pick up the prescription, but every cent the insurance company pays, we are paying by proxie with added admin costs.

YouTube offers an ad-free subscription called YouTube premium. It's reasonably priced and includes access to YouTube Music.

I like the hybrid approach of being able to be ad-supported or paid with no ads. I would like to see more of it.

What I don't like is a paid service like Amazon Prime that also includes ads. They include ads in their search results and they include tons of ads in their video library.

FWIW: Hulu offers paid access to content with ads but offers an upgrade to get rid of most of the ads, so there seems to be a whole lot of testing what works in this area going on right now, which I see as a good thing, I just hope that once everything settles the predominant model will be fair and respect user privacy.

Granted, all of that is true, but GP specifically differentiated between ASCII and ASCII Extended, then GP went on to say that after choosing the ASCII option and pasting the text in a text editor on Mac it was reported as UTF-8, which I was pointing out would be true because if the ASCII option is chosen as opposed to the ASCII Extended option then what he ends up with (ASCII) is valid UTF-8 as reported by the text editor.

Disclaimer, I just pulled this quote from Google ai which probably took it from somewhere else, but I just wanted to provide a little context. ASCII encoded text is also valid utf8.

> The first 128 characters of Unicode, which are the same as the ASCII character set (characters 0-127), are encoded in UTF-8 using a single byte with the exact same binary value as their ASCII representation. This means that any file containing only ASCII characters is also a valid UTF-8 file

I love python f-strings. I dont use the format specifiers that this article points out.

Also, even though use in log messages is discouraged, I go ahead and use them. It will let me know if there is some code path where the proper variable is never set. This usually comes out through testing, especially during fuzzing so I guess it really only works because of my testing, otherwise it would come up during runtime...

Thank you for the question.

Those platforms seem to be web analytics platforms, purpose built to track and analyze data from visitors to your web site.

Delve, on the other hand, concentrates on ingesting large quantities of structured, semi-structured and unstructured text based data. Think log files, csv, json, xml, item descriptions, etc.

The closest product to Delve would be something like Splunk or ELK.

I would say Delve is easier to get started with than Splunk or ELK, but actually offers more control since you can easily write your own search commands and field extractions. Also, role based access control is implemented to give fine-grained access to data.

Also configuration is all in one place, backend databases are configurable (sqlite, postgres, MySQL and any other database that works with the django ORM) and more.

Please let me know if you'd like any more details.

They really are.

Neither. I'm using its English definition:

To research or make inquiries into something

Good point. because of this and a few other comments, I am planning on renaming the project soon.

Any ideas you would like to share?

I do have a video I made a while ago, but I dropped the ball on starting a youtube channel to publish it.

Any recommendations on where I can upload and share a video, preferably one that I could embed in my README.md?

Thank you. I appreciate you pointing this out.

It seems like Delve is a popular name for software, which makes sense.

I am officially looking for a new name for this software.

It used to be called Flashlight, as in "A portable tool to shine some light on your data", but I liked the name Delve and a naive google search didn't reveal any projects named delve.

As an aside, Delve is not only popular in software naming, but it also appears in the English language writing quite a lot.

Thank you for letting me know. I did not know about the Delve MS product.

I just recently renamed from Flashlight to Delve.

I like the name and MS retired it, so I'm going to try to stick with it.

I am working on a versatile and powerful platform for ingesting, transforming, and searching through structured, unstructured, and semi-structured data. It allows for interactive searches, dashboards, alerts, and more.

We are in open Alpha at the moment, but plan on offering affordable plans while keeping the source code available. While in open Alpha and during the upcoming Beta, it is free to use for any purpose.

We will be selling Fair-Source license, meaning that the source code will be released under MIT 2 years after release.

Check it out here: https://github.com/DelveCorp/flashlight/

Feel free to ask any questions here or open an issue in the repo.

That's a good portion of what we are going for. I want it to be as simple as possible for users to make use of their data.

Whether the data is coming from logs, APIs, files or any other source, it should be simple and straightforward to onboard, search, correlate and visualize.

Once, the data is in Flashlight and you have crafted a couple of searches, then it is easy to put that search into a panel of a dashboard and share the dashboard with coworkers.

So, in essence very similar to both Splunk and Datadog, but there are some important distinctions as well.

* Flashlight will be much more affordable.

* Flashlight is self-hosted, so your data stays safe. Built-in and custom permissions also make it easy to share information with only those people you want to have access.

* Flashlight allows you to gradually model your data. Events are a nice, generic text-based data model, but once your data has special requirements, you can break out that data into a seperate model, which will allow you to specify only the specific fields you need, store the data in its own database table, apply custom permissions, generate CRUD API for the model and much more.

* Search commands, in Flashlight, are meant to be a lot more versatile and extensible. For instance, I have a custom Flashlight search command to connect to a group of servers, pull backups and store those backups for later. There is a corresponding custom search command to restore a specific backup later, so I was able to use those search commands in a "System Backup" dashboard with buttons to retrieve or restore backups.

Please let me know if you are having any issues with Flashlight by opening an Issue in the Github Repo.