Inject Your Preferences: Apps and Live websites

[Michael]

Yo,

Share what apps (right now I can do android, but learning) and websites you want live for a more or less permanent basis?

I've just spent a little time figuring out GitHub for the first time, easier than I thought.

That means, I'll be able to keep websites live indefinitely as long as GitHub remains available in a much more stable condition. 

Which one's do you want? I'm going through a phase where I'm testing out the coding abilities of AI, and with that, app and live website generation : now is the time to make requests, as it will only be for a few weeks or so most likely. So whether its the number series game, or some other, please let me know what your preferences are regarding previous applications I've made.

No guarantees, but feel free to share to share ideas as well across the spectrum cognitive training games and otherwise, for example just to encourage others, I just created my own personalized journal suited to my customized needs : for example, I have tabs that are sorted relative to my model of the mind and how my workflow streamlines rather than having to fit into some pre-furbished map that I have no control over. That's pretty cool for me, I've even added in a drawing entry option to every tab. My process is streamlined where I go from "Thought" to "Experiment" to "Adaptations", and I find that works pretty well for me. 

Best,

Michael

[Arnav Latkar]

I don't know. I want to be a genius though. In whatever way is possible through online training 

--
You received this message because you are subscribed to the Google Groups "Dual N-Back, Brain Training & Intelligence" group.
To unsubscribe from this group and stop receiving emails from it, send an email to brain-trainin...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/brain-training/f2b011f1-031e-446c-af76-f73b6f274050n%40googlegroups.com.

[Michael]

Genius really depends on four things (1) your own personal definition (2) your time in history (3) societies definition, which has only moderate value which is why it's not the only thing (4) sciences definition, again, moderate to high level but not total because it's seeking economically optimal understanding.

I would say, just my own 2 cents to give you some ground to work with, in today's age, genius is all about "how". Not "why" that dominates philosophy not "what if" that dominates sci fi, not "what" that dominates science, but "how", that's what more than anything else marries those 4 worlds. My answer is based on the fact that we live not in the tech era but more precisely, the AI era, and genius then, is about maximal leverage of this tool, which doesn't come from pure convergent but integration of lateral, divergent and synthesis based reasoning.

So, that's my 2 cents, with one principle that's primary there, and that is that as much as science is following an economic based optimisation function, so are our brains with respect to how it's training how we should process the world around us, which is why theres such a relativity to the end product of "genius", despite latent potentials. Another reason why my perspective integrates at the nexus of those 4.

My questions for you, are what does genius mean for you? And proportional to the amount you have that dream how can you ground that in a practical reality whereby you're still only led by your own vision which isn't dependent on the need for the approval, admiration or acceptance of the outside but instead by, simply appreciating and making the most of your own humble existence like all the rest of us, that's the secret to remember for our own unique paths for whatever we make of our own personal endeavours here.

That said, after you have answered this question for yourself, what are your ideas for how you can achieve this starting today? Doesn't need to be a grand gesture, but one gesture, leads you down your own path of independence, the sovereignty we are all gifted from birth, another practical reasons for looking after how our minds processes the world.

[Michael]

What I've decided to do regardless is create a cognitive benchmark site.

What I will do as well is reverse engineer latent 'process' of any one individual benchmark as a new take for how a user will reflect on any training they undertake on their default chosen training (i.e. meta-modelling).

I will create the github site by Sunday of next week, after which I will remain open to implementing improvements.

I got GPT5 now, yeah Claude just interestingly totally blocked me out and refunded my money at the same time they blocked me without my requesting, still pending feedback from them on the why heh! I'm a little proud of the fact that GPT5 fails miserably as well at improving not only my latest technique but just either (1) any of them (2) or coming up with anything that's truly original or even remotely close to being as good. Personally, I'd prefer it if it did come up with far better adaptations, but the over hyped reality of GPT5 despite what will be a site with cognitive benchmarks that will be easier for it to make, is just not the reality we live in, unless we take GPT's past stronger hallucinations as an analogy for accurate perception.

I'm not sure what benchmarks I'll have:

1. Backwards forwards digit span n-back makes sense

2. Basic number series n-back maybe

3. I should probably just include the most rudimentary digit spans as well because people love traditional measures as it's what they know?

4. Feel free to drop your benchmark or ideas in general before or after 

[Michael]

Draft: Cognitive Benchmark Site (GPT5).

I'll have to think pretty creatively between now and next Sunday, because GPT5 as I'm just now finding is just super, super good at re-creating pre-existing ideas. Genuinely impressed with its ability to implement pre-existing ideas, AI creativity will obviously take more time in my opinion. 

Try it out, all I simply asked GPT5 was this precise prompt "in HTML create a website that's designed to test a whole range of cognitive benchmarks and finish with an aggregate IQ score". 

And here was the result, again a reminder that only for a test run I've put it on codepen, the final site will be placed on a GitHub link that will always remain available for the end user for as long as GitHub remains accessible. 

https://codepen.io/qvmbioip-the-typescripter/full/ZYbJKjx

Between now and next Sunday (not this coming Sunday) I'll have to generate some new looks, at the very least decide whether or not to have the final 'aggregate IQ score' it spits out, which may or may not be accurate, I haven't put it through proper research filters yet. Best.

Rock out

[Michael]

"What I will do as well is reverse engineer latent 'process' of any one individual benchmark as a new take for how a user will reflect on any training they undertake on their default chosen training (i.e. meta-modelling)."

Aka what I'll be doing that hasn't been done before is not only mapping benchmark to hypothetical process (I.e. meta-modelling) but modelling each executive process to psychological outcome. This is the meeting ground between operational objective and methodological objective or fluid objective and executive objective. Very simply, it enables a user to generate more meta-feedback during the their training session on whether or not they're genuinely trying to max out their output. When we don't have many or sometimes, zero leads on feedback and we're just relying on our intuition, false confidence or just false measurements by our self-awareness are just naturally more common. For example, regarding the reaction time test, "Am I responding to right information as quickly as I can?" vs stroop test, "Am I noticing and responding to logical or stimulus conflict as quickly and as accurately as I can?". 

On the surface, this may seem pretty rudimentary, however when we measure what's going on underneath the hood and compare it with our psychology, we often mistake psychology with self and self with the experiences that are just simply biological outcomes, allowing distortions here far more likely to run rampant with Donald Trump being a great example here, despite many of his strengths. Thus, meta-feedback will act as a modelling between benchmark cognitive test as employed like the example site in the previous comment and self-monitoring of their same analogous operational objectives in the training technique (i.e.  meta-modelling), thereby creating an opportunity for the intersection between executive function self-monitoring and technique to be leveraged for max output. 

Breath-back variant of meta-modelling compared to Tiktok alone for example (see other thread to catch the line on meta-modelling), generally both requires and facilitates self-awareness, by not only being aware of the abstract mapping between measured benchmark cognitive process and self-monitoring of the optimal recruitment of underlying function their in training, self-awareness by its very definition is much less likely to self-hallucinate performance like a GPT does and instead make adaptations converted into meaningful gains within training sessions. That said, this is why it makes sense to be prudent about the composite of benchmark tasks employed so that we achieve that best conversion rate between abstract mapping between task and self-monitoring process. So again, either before or after "Sunday-Week", share whatever other tasks you can think of, take a gander at the cognitive benchmark site I made with GPT 5 just previous, and we'll move things to completion after this time and have something comprehensive with respect to both task and abstract mappings irrespectively. 

Thank you for your time, best wishes.

[Michael]

[ EDIT ] Dang phone.

Be as imaginative as you want.

I'm right now creating a syllogistic reading span test for example.

How I came up with this was by reflecting on when I trained on syllogisms a lot more some time back. One observation I made there was just the insane reading speed I had back then. I kid you not, I was reading and comprehensively synthesizing a book a day when I was really rocking it.

Lastly, what I'll be tweaking between now and the first iterative release is an aggregate self awareness score rather than an IQ score.

That's a lot more practical.

I'll promise to be more imaginative between now and the first release but it's totally true, with one caveat being that I trained with the associative pairing I've taught in the past, so not in the usual way.

[Michael]

Case Study Experimental Example - https://codepen.io/qvmbioip-the-typescripter/full/qEOXXXX

Let's see where it is by "Sunday-week". Otherwise thats it from me. As one self-awareness metric I add to the TikTok variant of meta-modelling, I only allow myself "one breath per video", any more time and I'm not placing enough pressure on myself as I'm more likely to overestimate how much time I need rather than underestimate. Moreover it's my backdoor self-awareness trick to get a reading on precisely how I'm performing in precisely the way I'm designing the performance metric of two posts ago. Intersecting reading span with syllogism span as an experimental progress metric with meta-modelling, is a solid, solid move in light of the progress I formerly made as expressed.

Let's see what the before and after will be

Best

[Michael]

Not words I never thought I'd express, but TikTok, albeit I am a new user, actually made me tear-up while meta-modelling. Its because you're analysing things on such a profoundly deep level, and pushing further, across parallel processing stretching however long your n-video-back is, picking up on and generating patterns on life and humanity at large. Combined with the automatic continuous self-questioning upon personal biases, inner conflict-stimulus resolutions and generalised perception, existentially, your identity has no choice but to evolve relative to the depth of the "complexity catharsis" that's achieved. Certainly one of the sub-goals of the action, it reflects high synergy as well between, cognition, emotion, identity and reality resolution routines. Haha. Peace out

That's a big hint to those interested in taking their training to the next level, start the learning curve today on meta-modelling. Rock out

[Michael]

Testing out GPT5's new 'agent AI' capacities.

To conclude in advance, this is the kind of task I'd be looking to incorporate as one of the benchmarks as a means of priming the right circuits in their personal learning phase of meta-modelling. 

Here's the best training task it could create in the mapping between relational reasoning and fluid intelligence and its subsequent training.

Any tuning that needs to be done will be completed by "Sunday-week" - https://codepen.io/qvmbioip-the-typescripter/full/dPYzJPW

An educational read below.

Here's the paper that the agent wrote.

https://chatgpt.com/s/6896da2c8f44819186ee2b0b3858562e

[Michael]

Yeah (see just above), this will likely be the best benchmark test and as stated, primer (or otherwise an actual genuine cognitive training task for beginners as some will certainly use it as, and that's perfectly fine of course).

We're all very lucky, this is great.

I'm forcing a user to choose 'match/no-match' as a way of forcing constant engagement. Learning curve will happen faster.

 Rock solid. 

[Michael]

I'll continue to improve any variant as I go along but please, share any errors, improvements however creative. Thank you!

[Michael]

Update/Repost ----> Apex will be ready to go in a few hours, more tweaks needed:

After being requested to provide some feedback on this variant, I have now created a permanently live site for Apex (with new updates) - https://ailivinlove.github.io/Apex/

You can just sit back, relax and let your imagination wonder now rather than having to stare at a screen. Moreover, I express this with the newly applied update of adding "Pause between relations in the same premise (ms): 6000" (a really great feature) as a customizable option for users now. I realise that before I should have added it, well, now I have. Please let me know if you'd like me to increase the customizability in any direction including adding more range to that timer. 

Let me know if you can spot any errors and I'll get right on it!

By the way, if anyone's looking for inclusions or adaptations to meta-modelling without a doubt the all in one package is definitely Apex 

You see I've thought about it a lot, and the only thing to really include that genuinely amps meta-modelling is cross-contextual reasoning and negation/contradiction. But they're so difficult to keep up, the fact that Apex has them already built in and there's only advantages to associating generated stimuli with the symbols of Apex, makes it not only a worthwhile inclusion but arguably a better or complementary primer to the matrices like relational reasoning game (which still needs to be improved in my view by the way) shared a few posts back. 

Best wishes, please share your own ideas with complete freedom, there's no need to hold back I'm open to anything and we'll soon have a full suite that's compact, organised and ready for deployment for everyone and anyone to use as a completely open source project.

Cheers!

On Monday, August 11, 2025 at 6:16:16 PM UTC+10 Michael wrote:

By the way, if anyone's looking for inclusions or adaptations to meta-modelling without a doubt the all in one package is definitely Apex 

You see I've thought about it a lot, and the only thing to really include that genuinely amps meta-modelling is cross-contextual reasoning and negation/contradiction. But they're so difficult to keep up, the fact that Apex has them already built in and there's only advantages to associating generated stimuli with the symbols of Apex, makes it not only a worthwhile inclusion but arguably a better or complementary primer to the matrices like relational reasoning game (which still needs to be improved in my view by the way) shared a few posts back. 

Best

On Monday, August 11, 2025 at 5:52:17 PM UTC+10 Michael wrote:

After being requested to provide some feedback on this variant, I have now created a permanently live site for Apex (with new updates) - https://ailivinlove.github.io/Apex/

You can just sit back, relax and let your imagination wonder now rather than having to stare at a screen. Moreover, I express this with the newly applied update of adding "Pause between relations in the same premise (ms): 6000" (a really great feature) as a customizable option for users now. I realise that before I should have added it, well, now I have. Please let me know if you'd like me to increase the customizability in any direction including adding more range to that timer. 

Let me know if you can spot any errors and I'll get right on it!

Best wishes, please share your own ideas with complete freedom, there's no need to hold back I'm open to anything and we'll soon have a full suite that's compact, organised and ready for deployment for everyone and anyone to use as a completely open source project.

Cheers!

[Michael]

God that was annoying heh, finally got it going, range of voices to choose from. Best. 

[Michael]

Voices need better alignment (male/female mixup in some instances) but overall not bad for Apex, otherwise I've changed time per trial to time per premise, I thought that'd make it flexible and easier. 

Any adjustments let me know, however practical the better, however creative though... the mind-blowing. 

Thank you for your time.

Best.

[Michael]

1. Logic of matches concerning Apex corrected. 
2. Will make all cognitive tests available on one GitHub hosted site 
3. Updates will still arrive until perfected following posting of first full draft

4. Will be continue to be open to alternate ideas / citing of errors

5. Will likely repost this thread after the first draft of the site has passed enough improvement feedback loops. 

Thanks!

Rock out

[Michael]

This link will be updated as we move forward - https://ailivinlove.github.io/CogniVolution/

[Michael]

Edit: Sound for Apex does not work on some desktop/browsers, if it's ever an issue you can't get around, just simply share the issue and I'll update the code. But because it doesn't effect me and no one has mentioned anything, I'm leaving as is for now.

Apex now up and running, GitHub/other minor issues - https://ailivinlove.github.io/Apex/

Please report any glitches, errors, bugs, etc. Thank you! 

Rock out.