The basic error underlying this line of thinking was covered above, but I will reiterate. And at risk of sounding cliche, without taking the holistic picture into mind, the only thing being observed is a different part of an elephant.
Statistical observation against a greater sample is flawed. That would necessitate that there is an ideal state of human functioning (mentally/physiologically/etc.). While I'm certain there is a local ideal per individual, and that many features for an idealized "perfect state" are shared among many (e.g. proper diet, rest, etc.); psychology is not so simple. Again, dysfunction in the brain would necessitate there being an ideal state of function -- which has not been defined or determined outside of individual traits that have been interpreted to be at odds with various (subjective) ideals of being.
I do realize that certain physiological states make it subjectively harder to exist in one's specific environment. My issue is with taking this as given fact, rather than interpretation -- that there is an abstract and all-encompassing ideal to be reached, regardless of the individual.
This is ignoring the erroneous usage of medical measurements for purposes of "matter-of-fact" and not recognizing that measuring physchological behavior, as by-product of the underlying nervous system and all the other bodily systems, is correlative. To use a simplified example: if you take an EEG of an "ADHD" brain against an average sample of "non-ADHD," yes it will look different. If I were to steelman it, I would swap out "non-ADHD" for those with prefrontal cortex brain damage. Now the results may look similar. However, this doesn't extend to reaching the conclusion that someone with ADHD has brain damage (they may -- but the results are not causative, merely correlative).
Or what I've already written above, that the data measured is the right data to measure.
The environment, the people, their norms, behaviors, and "average" are different depending on the specific geography, and constantly changing. To nail down an "ideal," in spite of the specific circumstances of each individual is -- in my opinion -- misguided. For an off-hand example: a wealthy and erratic person is an eccentric; but a poor and erratic one is clinically insane. Or someone with hallucinations in Western society is a schizophrenic; while in other cultures he may be a shaman.
I want to take a tangential excursion into the various cultures and their implications against an "ideal" a la Quigley and Huntington -- but I've already written too much.
I'll step in for rozal, because this is a big ask (and much more effort than picking out a few research papers, that no one here will surely read).
/*********/
The main issues with ADHD (and the current crop of mental health disorders) are postulates that researchers, practitioners, and regulatory bodies (RPR) all take for granted, without rigorous inquiry:
1. There is a "normal" state that some majority of people are in
2. There is an "abnormal" state that some minority of people are in -- with various subcategories/subsets of "abnormal"
3. That the "abnormal" set of people requires intervention to help move them into the "normal" set
4. That the set of people currently proposing and executing interventions (RPR) are the right people
5. That the current set of methods and tools used by (4/RPR) to move the "abnormal" to the "normal" is the right set
6. That (5) achieves the right goals
/*********/
For #1, this has never been formally agreed upon. There are various interpretations of what a "normal" life is, depending on one's own: value system, culture, upbringing, current environment, professional training, etc. Likewise, this "normal" may take on other labels, depending on who you ask: e.g. "fulfilled," "actualized," "happy," "content," "productive," "well-lived," "spiritually-rich," "good enough," "tolerable," "economically stable," etc. In lieu of any centralized and concrete definition, RPR has unconsciously gravitated to not defining what's "right," but what is "wrong."
///
For #2, there is again no concrete and unifying criteria for what satisfies "abnormal," but a constantly changing set of qualifiers based on: statistical frequency (a), executive function (b), adherence to an ideal set of traits (c), adherence to social norms (d).
In (a), the basic idea is that statistically prevalent traits are the "norm" and those less prevalent are the "abnorm"; i.e. the mean set of traits is treated as an ideal. Generally, this method makes no subjective judgement on the nature or "rightness" of the traits, just that they are the most prevalent. Readers can generalize from their own lives, times when groups of people exhibiting similar traits were not what one would consider "normal," by personal definition.
In (b), executive function* concerns one's ability to self-regulate, deal with stress and adversity (i.e. adapt), and participate in one's society, i.e. sustain one's life as a human being involved within a group of other human beings. Once again, no subjective judgement is made on the nature or "rightness" of the environment or group one find's oneself in -- only that the individual is able to meet the demands of his or her environment. Readers can generalize from their own lives, times when environments are sufficiently harmful to themselves that adaptation to, rather than escape from, such environments would've been considered in their minds "abnormal."
* as reductively, but generously, rewritten by me, paraphrasing the gist of Rosenhan & Seligman, while ignoring the obvious subjectivity around "rationality," "moral/social standards," and "appearances/how one appears to others"
In (c), we start to get the first concrete, yet not-yet-popularly-accepted, definition of a "normal"; paraphrased from Jahoda (a la Wikipedia):
- Efficient self perception
- Realistic self esteem and acceptance,
- Voluntary control of behavior
- True perception of the world
- Sustaining relationships and giving affection
- Self-direction and productivity
Plainly, many of these are highly subjective, and dependent on the environment one find's one in. I will not be digging into these, because I too have my own biases, values, and so on that would color these differently. However, these will be relevant to (4).
In (d), this is self-explanatory: does the individual adhere sufficiently to his environment's social norms? Once more, no subjective judgement is made whether or not the norms are "right," only that they are the prevailing ones.
///
For #3, the unspoken justification of providing treatments to the "abnormal" is that it brings benefits to the "normal." Reductively, one only carries out tasks that benefit oneself. These could be financial, social, spiritual, or otherwise, incentives that move one to act for some imagined benefit.
One could make a case for allowing the local community to assist the "abnormal" to "normalcy," because resources are limited, and each member relies on the others to survive and thrive. This can be seen in tight-knit farming communities, where no single individual can realistically survive alone; and the addition of this individual into a beneficial member of the group allows the group access to more resources, ergo increasing its chances of prosperity.
On a more national level, one could make the same case for assistance that is outside of the community, but within the overarching "group" that encompasses all the other groups. For example, institutionalized mental health treatments will allow the populace of a country to contribute to the country's goals, same as they would to their local community.
Now, once again no subjective judgement is made on whether the goals of the local community or the goals of the nation are "right" -- or any philosophical quandary raised about the necessity of adhering to these goals. Readers can decide for themselves what extra-personal goals are useful to their ends.
///
For #4, there has never been any widely popular questioning of who should be "right" ones to assist in the "renormalization." Simply, there was a power vacuum, and people (RPR) filled it in to achieve their own ends.
///
For #5, yes there have been a plethora of studies and experiments, and so on, to test efficacy of treatments towards various fuzzy ends (e.g. a reduction or increase in certain subjective traits, experiences, and criteria, or that of more objective economical figures). However, there has been no popular questioning of whether or not the tools (medication and the current flavor of physcological therapies) are the best ones to be used.
In the U.S., the institutionalized mental health systems have a near-monopoly on such services. Partly due to their academic roots, all knowledge must be built on what has already been "discovered" and deemed "right." Anything not aligning to the past, will not be given due consideration as legitimate -- thereby artificially restricting the global maximum set of available tools, to a local maximum set of "popular" tools (regardless of their actual substance).
///
For #6, we look again at #5, and ask whether the "fuzzy ends" are the right ends to be sought after. No popular discussion has been made here.
///
In fairness, "right" can only ever be a personal definition, based upon one's own: values, biases, culture, upbringing, etc. -- and others can either agree, and spur on the legitimacy of a certain definition of "right," or disagree and lower its power.
The main question to be formulated with all this background information is: "are the current treatments for mental abnormalities the best for the individual, or the best for the whole?"
With consideration to all the ideas I've laid out in this post, I believe that the current treatments for mental health abnormalities greatly benefit the larger "whole" rather than the "individual." Medications come with a plethora of side-effects, and potential long-term issues. Likewise, the long-term efficacy, sustainability, and outcomes of behavioral and psychological treatment via therapies are under question by me. Entire livelihoods and industries are built upon the treatment inherently (i.e. institutional mental health services), but also its after-treatment effects (i.e. "you can function at a job, and help grow the economy").
My belief on "right" is to bring about a life to the individual where their mental health abnormalities do not require constant reliance on external forces to allow the individual to live the life that best fulfills his or her own needs. Otherwise, the incentives are grossly weighted towards the external force's ends, rather than said individual's.
///
N.B. Whether or not ADHD is legitimate is a subjective discussion -- and one that I believe misses the point. I think it can be further reduced down to: "there are people with a different set of traits that institutional players have decided to umbrella under 'ADHD', with all the various treatments involved. Is this the right way to go about things?"
We both know that's not the most charitable, good-faith interpretation of my comment.
I think I've made it clear that I have personal experience in this area; but if it wasn't, I'll add in a disclaimer: I've been diagnosed with ADHD as a child; I have taken and continue to take stimulant medication. I've been through the ringer, passed around from one mental health professional to another -- along with personal research and experimentation.
I do not like stimulants. I think they are useful for masking unresolved underlying issues, where necessary. Other times, I detest their usage as PEDs in the workplace.
I think there are better ways to "treat" ADHD. Unfortunately, many times one does not have the resources or information to do so well.
Currently I'm working through CPTSD treatment to resolve what is essentially a constant inability to relax or focus, or maintain a "productive" and "well put together" life, or not self-destruct every time things take a poor turn. If you're being treated for ADHD solely by stimulant medications, I would urge you to explore further.
- (mis/ab)use of drugs to artificially increase work output (with lower quality), causing the rest of the team to have to: fix the (mis/ab)users mistakes, and deal with the political fallout of appearing to be slacking relative to the (mis/ab)user
- mismanagement of prescriptions, either by the prescriber or the prescribee: very easy to slip into psychosis or not understand that you shouldn't "feel" the drugs, and that you don't necessarily need to increase the dosage just because the initial high wore off. If your life is being held together sufficiently well on your current dosage, and whenever you forget to take your meds your life falls apart, then your dosage is perfectly fine. Easy to not realize your sleep has been severely impacted, but is being masked by the stimulants, i.e. you're riding on the edge of "psychosis."
- disregard for lifestyle modifications that synergize with ADHD and medications (e.g. good sleep hygiene, better diet, regular exercise, meditation, etc.)
My close(ish) confidants in my industry will be understanding; and those outside of my inner circle will be polite and cordial about it; but behind closed doors people will judge you, and use it against you.
Ask your manager for a waiver, and an invention assignment on your side-project.
It’s not a business; but a side-project (right now) — assuming it’s tech-related. If it’s a non-tech business, where you’re not generating IP, even better.
> What can I do to minimize my risk of having lower chances on the job market due to this adventure?
a. Don’t quit your day-job until the side-project makes atleast two times your current comp.
b. Work on it for a year; realize you’re way in over your head and pretend it never happened. Line on CV: “took a sabbatical for one year.”
> What data do recruiters and employers look at to determine if the failed startup listed on a CV is a plus or a minus?
Most recruiters will not understand. Data? The data is that you were essentially unemployed for a long period of time; and instead of breaking out of the mold, you came back to the fold — and so you will be judged accordingly.
Employers? Most will think you’re at best a flight risk and at worst too independent-minded to be an employee.
Some will view it as a unique plus if you can show: a. A product; b. Paying customers; but these are definitely the minority and you cannot plan for finding these people. If you do decide to drop out of the corporate grind, your best way back in would be your network. If you don’t have one, you may end up working for boiler rooms at much lower pay.
No, reducing LLMs to just “statistical text predictor” is an absurdity; but so is leisurely stringing together observations from unstructured experimentation to form a “matter-of-fact” conclusion.
Read the ChatGPT paper and specifically hone in on “transformer” and “input processing” (paraphrased). It will give you a clearer picture of what it actually is, rather than appears to be.
2. Owning a serf rather than establishing an engagement with a professional
On #1, older, more experienced workers are not going to accept bottom-of-the-barrel compensation for the outsized impact they make relative to. Generally, this is just "free market economics" looking to maximize returns compared to costs. The easiest way to get costs down is to keep payroll down (in service-based businesses, as many tech and tech-reliant firms are, payroll is going to be your highest cost). It's low hanging fruit; and by the far the easiest way to increase the bottom-line (see: startups).
On #2, this is mostly more relevant in large, established companies. The main goal on everyone's mind is no longer to make money, but to climb the ladder. Yes, costs and profits are important, but only in times of "bad vibes." In better times, everyone is trying to justify why they're a part of the gravy train, and to get a bigger portion. Usually, this is achieved through utilizing "human resources" to work on "big projects" (that usually do not materialize anything of note -- but generate impact, i.e. intrigue ). In other cases, a newly-minted lord will not be given sufficient bodies or budget to expand his realm, so he'll need to squeeze as much as he can out of what he has in order to make do. Older workers are less tolerant of bullshit and being squeezed for no material reward to themselves.
Generally this means that older workers will not be the first-pick for startups (cost-conscious, "scrappy," i.e. #1) or well-established bigcos (reliant on transmuting human souls for the enrichment of the faux-nobility, as in #2). In either case, neither type of company is looking for great technical talent, but just a body to achieve an end. Any schmuck can fill either role -- because competence isn't necessary for the success of whoever controls the purse strings. Unfortunately, these two types of companies are the prevailing sort; thereby "ageism" can be seen in tech, the same way educational box-ticking has become en vogue: it is politically more advantageous.
Companies with sane leadership, and bodacious-yet-grounded vision, that are looking to actually offer something of substance will value competence over anything else (look at old SV startups for an example). But once again, they're an unfortunate minority, and they most likely will not be doing cold recruiting (as it is insane).
Middle: ORMs relieve the impedance mismatch of the OOP "object graph" paradigm and the SQL "relational algebra" paradigm. One could even make the argument that this mismatch is inherent and we're just doing CPR on a rotting horse -- so now the industry standard is to use NoSQL databases, such as MongoDB, to get away from bygone ways of thinking about data access. Likewise, SQL is not conducive to thorough testing coverage; and migration management is prone to user-error. In this vein, the industry has once again innovated and revolutionized data-access, moving away from "CREATE READ UPDATE DELETE" to "ENCAPSULATE INSTANTIATE INJECT CREATE READ(+UPDATE|DELETE)* TRACK PUSH SYNC." With all this in-mind, it would be utterly baroque to use anything other than JS+Node+Mongo+Mongoose for a fully unified front-end and back-end.
*An astute reader will recognize that ORMs are constantly doing N+1 queries (to first pull the data, transform it into objects, update said objects, and then push the changes to the database). We feel as though hardware has gotten sufficiently advanced that the mind-space these costs inhabit is no longer justified -- and it is OK to do things this way.
Yes, but that would require academics to grow a practical bone in their body and realize "understanding" language is just one piece of the "intelligence puzzle."
The hype will die down on LLMs (slowly... all those ML researchers need to justify the sunk-cost of specializing into a very niche, albeit sales-friendly, field).
If it's any consolation, I doubt we'll see progress towards a "technological singularity" until the current crop of career scientists retire into the dirt -- or there is a fundamental change in resource allocation allowing other, more creative types to start experimenting with building out a probabilistic model for:
Human language in -> determination of "what to do"/workflow to run -> run
We're seeing a few startups in this area, but I haven't seen anyone create any useful agent.
It's for when you want to have distributed data stores.
For example, if you're building out a doc to track your net worth, income, and spending habits:
- One main sheet that collates all the data required here (e.g. credit card statements, in-flows and out-flows from bank accounts, current debts and current assets)
- Have 3 more docs:
One that collates and summarizes your credit card charges
One that collates your bank statements
One that collates your assets and liabilities
///
Let's say my "credit card charges" sheet automatically pulls in my charges over API, and appends them to the "Charges" sheet in the "Credit Card.xls." Then the "Summary" sheet in this "Credit Card.xls" summarizes this information into something useful. My "Main.xls" (that collates all of my data into something even more readable and useful) can then pull data from this "Credit Card.xls" sheet (through API or locally), and automatically keep itself updated.
I could stuff this all into one single doc, with numerous sheets, but I don't want to deal with the cognitive overhead of having to navigate through an enormous amount of sheets I rarely (if ever) need to touch (again).
Perhaps I don't even have access to the physical "Credit Card.xls" doc, because my personal assistant automatically appends to it, and keeps it uploaded somewhere.
Use a view that acts as a quick-and-useful abstraction to mimic a denormalized table?
E.g.
```
CREATE VIEW vw_events_and_projects AS
SELECT *
FROM events
JOIN projects
```
Then
```
SELECT * FROM vw_events_and_projects
```
Edit:
And if you need OLAP, replicate the normalized table to a database that handles analytics workflows better (e.g. ClickHouse).
Then you get the normalized forms for your OLTP workflows (your "bread and butter"); and you also get the efficiency and ergonomics of real-deal OLAP.
Of course, your biggest issue is going to be keeping the two in-sync. Obvious case is to have your OLTP database stream synchronization data to the replica whenever data is modified.