Data Disinformation: the Next Big Problem

Automatic code generation LLMs like ChatGPT are capable of producing SQL snippets. Regardless of quality, those are capable of retrieving data (from prepared datasets) based on user prompts.
That data may, however, be garbage. This will lead to garbage decisions by lowly literate stakeholders.
Like with network neutrality and pii/psi ownership, we must act now to avoid yet another calamity.

Imagine a scenario where a middle-manager level illiterate barks some prompts to the corporate AI and it writes and runs an SQL query in company databases.
The AI outputs some interactive charts that show that the average worker spends 92.4 minutes on lunch daily.
The middle manager gets furious and enacts an Orwellian policy of facial recognition punch clock in the office.
Two months and millions of dollars in contractors later, and the middle manager checks the same prompt again... and the average lunch time is now 107.2 minutes!
Finally the middle manager gets a literate person to check the data... and the piece of shit SQL behind the number is sourcing from the "off-site scheduled meetings" database.
Why? because the dataset that does have the data for lunch breaks is labeled "labour board compliance 3", and the LLM thought that the metadata for the wrong dataset better matched the user's prompt.

This, given the very real world scenario of mislabeled data and LLMs' inability to understand what they are saying or accessing, and the average manager's complete data illiteracy, we might have to wrangle some actions to prepare for this type of tomfoolery.

I don't think that access restriction will save our souls here, decision-flumberers usually have the authority to overrule RACI/ACL restrictions anyway.
Making "data analysis" an AI-GMO-Free zone is laughable, that is simply not how the tech market works. Auto tools are coming to make our jobs harder and less productive, tech people!
I thought about detecting new automation-enhanced data access and visualization, and enacting awareness policies. But it would be of poor help, after a shithead middle manager gets hooked on a surreal indicator value it is nigh impossible to yank them out of it.

Gotta get this snowball rolling, we must have some idea of future AI housetraining best practices if we are to avoid a complete social-media style meltdown of data-driven processes.
Someone cares to pitch in?

TL;DR

Got Salesforce customised by an external company, took over the code and wanted to cry.

So my company decided to get a new crm for sales with contract generation and a whole lot of fluff. Specs included an easy to implement API to connect to our in-house software and an easy to adjust contract Wizard.

After month of checking various companies the CEO finally settled on...
Drumroll
The cheapest German based company.
Turns out the only part that is based in Germany is the sales department, development is based in Poland which made for interesting times during implementation because of the language barrier that comes with non native english speakers.

We as in our development department for told that we wouldn't need to worry about the solution because it would all be developed and maintained by the new company. As we are fairly small that was more then welcome.

Fast forward to integration day. No docs for the API available, contract Wizard is hardcoded, bunch of errors and inconsistencies. Get tickets for them.
Ok we can deal with this. Just tell the ticket writer he needs to address the problems with the external. Yep, doesn't work. External fixes bugs and introduces then elsewhere.
Fml. Ok I'll take a look into the code. Ugh, Java, I hate that shit but at least I don't need to worry about all the fluff, just the code, so it's ok.
No repository present, code is developed in the Dev environment and pushed to prod. Ugly but works. Code comes with a lot of functions but only one real class called "CommonUtilities" ... even the web API is defined in there. Meanwhile my colleague throws out the need for the API because we will just directly tie to SOQL. I'll let's check what's going on here, nice you reused Lead/Account/Contact for the branch offices of companies as well...
Is not like Salesforce has a bunch of logic tied to these objects...
Nice the required implementation of an automatic holiday import is not even there, just a custom object populated with this year's data...

Tell CEO how badly all of this was handled. Nice note this dumpster fire is our new in-house project because CEO cancelled the contract with the external.

Ok we can deal with this, let's set up a repo, define the CI/CD and get the extensions for vscode. Nice now this all makes sense. Fix all bugs and reimplement the contract Wizard using custom objects that sales can change so the contract is actually easily customizable by a non dev. Implement branches as their own object to avoid opportunities, triggers and the likes to be executed for them.
Took a whole day. Why did these 2000 lines of code that was shipped to us take 3 month to implement?

Lesson learned: never trust an external to just do a good job.
New rule implemented by company to always have a Dev check in regular intervals on projects handled by externals for standards and overall logic

10k down the drain for what amounted to 2 weeks tops (one dev) if we had handled the implementation of Salesforce and all the requirements in-house

IT jobs explained with a broken light bulb 😂

nothing like building a lead generation site in the advertising sector and then running a test through the site with an addblocker turned on.

The next step for improving large language models (if not diffusion) is hot-encoding.

The idea is pretty straightforward:
Generate many prompts, or take many prompts as a training and validation set. Do partial inference, and find the intersection of best overall performance with least computation.

Then save the state of the network during partial inference, and use that for all subsequent inferences. Sort of like LoRa, but for inference, instead of fine-tuning.

Inference, after-all, is what matters. And there has to be some subset of prompt-based initializations of a network, that perform, regardless of the prompt, (generally) as well as a full inference step.

Likewise with diffusion, there likely exists some priors (based on the training data) that speed up reconstruction or lower the network loss, allowing us to substitute a 'snapshot' that has the correct distribution, without necessarily performing a full generation.

Another idea I had was 'semantic centering' instead of regional image labelling. The idea is to find some patch of an object within an image, and ask, for all such patches that belong to an object, what best describes the object? if it were a dog, what patch of the image is "most dog-like" etc. I could see it as being much closer to how the human brain quickly identifies objects by short-cuts. The size of such patches could be adjusted to minimize the cross-entropy of classification relative to the tested size of each patch (pixel-sized patches for example might lead to too high a training loss). Of course it might allow us to do a scattershot 'at a glance' type lookup of potential image contents, even if you get multiple categories for a single pixel, it greatly narrows the total span of categories you need to do subsequent searches for.

In other news I'm starting a new ML blackbook for various ideas. Old one is mostly outdated now, and I think I scanned it (and since buried it somewhere amongst my ten thousand other files like a digital hoarder) and lost it.

I have some other 'low-hanging fruit' type ideas for improving existing and emerging models but I'll save those for another time.

Somehow I found Rousseau's the social contract.

I'm early into it since you know fucking chomo faggots with no balls keep screwing the world up trying to steal real peoles personalities and make them queer which eventually will lead to a generation that murders them being bred.

Anyway I found a love phrase.

Slaves loose everything in their chains even the desire of escaping them.

He continues.

Force made the first slaves, cowardice perpetuated the condition.

In short
The world being full of cocksucking perverse house niggers that love the taste of table scraps is the problem of the free man whose life is being devoured by scum like tosensei

Hello all, recently I have been doing alot of front end work in web forms and lead generation. I would love to learn more about marketing and how it can be applied as a dev.

Does anyone recommend any good books atall?

Thanks!