Posts

Posted on

small year-end observation of GenAI/LLM/transformer

gotten my own evidence of how far GenAI, based on the Transformer model, is going (or is going nowhere) yesterday while finalising my last piece of homework for 2025. GenAI, based on the Transformer model, works fundamentally by predicting what word(s) come next. and what made this ‘prediction’ possible? the dataset used in the training that the models have gone through informs this. in short, the Transformer, while ‘creative’, is creating based on existing patterns derived from dataset. and who created this dataset(s)? human thinking, thoughts, ideas, formed into words in the pre-GenAI era. and that dataset has long runout by now. you may read this article by de Gregorio to see all the ideas i have mentioned fall together.

long story short, whatever LLM provides you, it’s something that existed out there in its mega training dataset.

so, now back to my observation. this is the statement i wrote/created:
“With the advent of generative artificial intelligence (GenAI), cyber actors have harnessed it for autonomising complex hacking activities”

after feeding the statement into PAIR (powered by claude), platform suggested:

“autonomising” –> “automate“ （clearer expression)

what’s clear, what’s not clear is subjective. but, “clearer” here is a conclusion of the algorithms based on the dataset. and why is “autonomising” less ‘clear’? by design, ‘clearness’ has to be interpreted based on its training dataset. begs another question, autonomising vs. automating/automate, which term is likely to appear more often in the dataset, and thus lends to the prediction of ‘clearness’? from my author’s point-of-view, PAIR’s suggestion is definitely not ‘clearer’ in representing what i intended for my readers. and, ‘autonomising’ is likely a relatively rare concept out there at the moment. to me, in this case, LLMs’ greatest limitation of being bounded by its dataset is somewhat revealed. asking a far stretch question, is the current conception of LLM/transformer going to lead to AGI? i think the answer is clear.

Posted on

of cats and guardians of staircases

how many cats and guards of staircases have you observed in your life? no? not sure? read on …

The Ashram Cat (aka The Guru’s Cat)
An esteemed guru (spiritual teacher) is teaching his disciples, but an ashram cat constantly distracts the students by wandering around. To prevent the distraction, the guru orders his disciples to tie the cat to a post or tree during lessons or evening worship. This practice continues daily. Generations of gurus, disciples, and cats pass away, but the act of tying a cat during the lesson becomes a deeply ingrained, sacred tradition.

The Guardian of Staircase
John F. Barker in Roll Call tells the story that for more than twenty years, for no apparent reason, an attendant stood at the foot of the staircase leading to the House of Commons. At last someone checked and discovered that the job had been held in the attendant’s family for three generations. It seems it originated when the stairs were painted and the current attendant’s grandfather was assigned the task of warning people not to step on the wet paint. (source acknowledgement: www.lecturesbureau.gr)

time has changed, ubiquitous network connections and mobile devices and apps have arrived, and GenAI has descended. but things done yesterday are continued today, tomorrow, and probably the day after tmr. who, especially minions, dare question ‘traditions’ or remove guards from anywhere?

Posted on

meaning-making, losing it?

kueh attended a meeting and heard his boss huat shared on his latest beloved AI tool called LnbkM, and how he used it to summarise the so many readings and articles on the internet. boss huat encouraged all at the meeting to do what he does.
as kueh listened, he scratched his head and felt confused. slipping his hand into his left pocket, he took out his recipe and flipped to some notes he jotted down just last week, when attending a talk by a professor How PL who shared the following (research-based) ideas abt human learning:

  • learning is about meaning-making. and reading is a means through which meaning-making takes place.
  • neuroscience (brain-based) research also suggested that cognitive functions may be lost if they are unused or used less — comprehension, analysis, synthesis to just name a few

what boss huat suggested was to let AI be the one to “read”, and humans only receiving the summary of “reading” by the AI. if so, who is doing the meaning-making? kueh wondered what would be the future of humans be like if they begin to lose the brain functions of meaning-making? if so, what does it mean to be humans anymore?

Posted on

rethinking meetings and decision making

what’s the purpose of calling (council/committee/team) meetings? decision making is one primary reason.

why decision making only during meetings? cos these decisions often need discussions to make meaning of best course of actions. 

why does “meeting” exist? cos pple who need to make decisions are busy with many other things in life, and meetings “guaranteed” and forced pple to set aside time, locked tgt in a space to discuss.

in the pre-internet age, this is absolutely essential. in the post-COVID19 age, workplace are already relying on WhatsApp and other messaging/meeting tools to discuss things and make decisions instead of only at “meetings”.

some held strongly to the idea of “decisions already made at meeting” = cannot be changed. why?

in the pre-internet era, there’s no way to conduct discussions without pple meeting (again).

in the post-COVID19 era, with group messaging apps such as WhatsApp, are discussions to enable meaning-making not possible? 

but no, the counter argument is cos meetings are recorded and noted (cf. minuted) and are “official”. in the internet era, digital records can be served as evidence in courts, are “meetings” necessarily more serious business than courtrooms? 

but again, no, cos face-to-face interactions lead to deeper discussions!? well, more dynamic perhaps; deeper, is questionable. cos meetings are time-limited and one hardly has time to draw additional references/resources to bring discussions deeper beyond that space and time defined.

a meeting is a gathering of humans, and thus human behaviours and psychology come into play. there’s always pple who are more passive and some who are more vocal at meetings. and gathering in a group hearing the more vocals can easily end up in group-think. why? it’s easier to agree than disagree, that’s the ‘lazy’ nature of brains, not to mention if one has some other things (family/workplace/whatever) on her/his mind. and no prize for guessing what’s the outcomes/decisions made in such “meetings”?

formal meetings are still needed. cos the society expects it, and there are decisions that would warrant deeper deliberations face-to-face dynamically. but the argument that decisions made at meetings are cast in stone and cannot be changed warrants a rethink as discussions are now enabled by technologies and “recorded”. especially for decisions made when the more knowledgeable others are absent and the rest are discussing “in the blind” if not less knowledeable.

if ideas can be improved and decisions can be improved outside “meetings”, why do some pple dearly hold on to meetings (and the decisions made at meetings) conceptualised by last century definition? well, mindset, beliefs, and values are the most difficult to shift (not impossible), not any amount of SkillsFuture credits would make it easier. something all too familiar to educators and researchers in the field of educational psychology and learning sciences (:

Posted on

only good and good-er

realised i have not posted to elaborate on the philosophy of ‘good-er’ although i have used the term in a few posts in the past. so shall do it now.

yes, it’s non-existent in English and a grammatically wrong word.
but guess that made the word stands out even more, cos in the ‘good-er’ worldview, there’s no right or wrong; only good, and good-er exist. the philosophy can be interpreted as an extension of the core of the 12 principles of knowledge building (Scardamalia, 2002; Scardamalia and Bereiter, 2006), i.e., improvable ideas.

often, in the ‘right vs. wrong’ world, it’s not uncommon that we hear: ‘this idea is stupid, it is wrong, it’s unacceptable’… and pple began to dismiss their own and others’ ideas and not to mention the emotional side of things, bad, sad, angry …
contrastingly, in the ‘only good and good-er’ world: ‘all ideas are good ideas!”. marrying it with kb’s principle of improvable ideas, all ideas become improvable. end result: all good ideas can and will only become good-er ideas over time.

the good-er philosophy was first instituted as part of my design/model for blended learning for CL teachers’ professional development. it was my research project between 2015-2016 when i was with the SCCL. it’s the first ‘rule’ (among three) that my community of teachers adopt: “只有好，和更好”。 so if u ever need to cite a source for the good-er theory/philosophy, you could either use this blog post (Tan, Y. H. (2024, November 9). Only good and good-er. Edublog.net. https://edublog.net/2024/11/09/only-good-and-good-er/ ), or my handbook if you prefer an earlier source:

Tan, Y. H., Tan, Y. N., & Chow, F. Y. (2019). Blended learning for in-service teachers’ professional development: Handbook for new instructors. Singapore: NTU-SCCL Press.
OR
陈育焕、陈雁妮、周凤儿 (2019). 混成式在职教师培训：新手指导员手册 [Blended learning for in-service teachers’ professional development: Handbook for new instructors]。新加坡：南大-新加坡华文教研中心出版社。

Nov 7, 2024, Co-generative dialogue on educational research @RGS

towards only good and gooder!

Tan, Y. H. (2024, November 9). Only good and good-er. Edublog.net. https://edublog.net/2024/11/09/only-good-and-good-er/

p.s. scan QR code for a draft of the handbook — Tan, Tan, & Chow (2019) — that documented the 2015-2016 research
Proudly powered by WordPress