LLMs, chatgpt & ubi 4.0

with the blazing speed that LLMs are (or commonly known as “chatgpt”) developing, some people may be worried abt job security.

it looks more like a case of “the rich gets richer, the poor gets poorer” 贫者愈贫,富者愈富, not in the monetary sense, but the knowledge creation sense. more precisely, LLMs/chatgpt are going to make the expert-layman’s speed/efficiency gap ever bigger. layman can only (blindly, unknowingly, ‘trustingly’) copy-n-paste without understanding (cos they do not have enough prior knowledge to assess the output), while the expert can build on what LLMs/chatgpt throw out and idea-improve repeatedly with the system with further prompting and/or data input.

based on the above theory, to ensure everyone’s livelihood and well-being tmr (not limited to those who are worried abt job security), the idea of universal basic income (UBI) would probably need to upgrade to at least a 4.0:

UBI – food, water housing
UBI 2.0 – food, water, housing, power/electricity
UBI 3.0 – food, water, housing, power/electricity, wifi
UBI 4.0 – food, water, housing, power/electricity, wifi, negotiate/prompt LLMs

so what would UBI 4.0 mean for a k-12 school teacher? oh btw, openai just released its Code Interpreter within chatgpt4 few days ago. another game changer?

and thanks ziwei for the early brain-waking convo (:

teaching home-based (online) learning

the last post was created on 0214, and today’s 0412. on 0214, the covid-19 situation was just budding, and now on 0412 the situation has escalated worldwide. on 0214, classroom teaching was still going on as per ‘normal’; on 0412, the new norm today (which began 4 days ago) is full home-based learning (or FHBL in short). so how has learning become, from one norm to another norm?

chanced upon NUS Prof. Ben Leong’s post in which he reflected on the his FIVE(5) years of journey before achieving success in conducting full online interactive teaching. epistemologically speaking, we know learning is slow and experiential, and i thank prof leong for sharing his (giant’s) experiences for the rest of us to learn from. these are some of the ideas/statements that resonated with me (with emphasis added):

  • “the first principle of leadership is that ‘if we take care of our people, our people will take care of us.’ Similarly, ‘if the students know that the teacher cares, the students will learn.'”
  • “My hypothesis for why online teaching has not been successful in the past is that teaching is an inherently social activity. Things are different today because we are now able to replicate many social interactions using available online tools.”
  • “To make learning social, I have successfully adopted and deployed the following: (1) Interactive online teaching using Zoom; (2) Activity feeds; (3) In-platform messaging; (4) Forums – regular and video; (5) Gamification” (numbering added)
  • “Students actually have to be taught how to be online students and to interact with the online platform because lessons can be conducted. However, once the protocols are established, it can become quite natural.”
  • “… the natural approach to making online learning more social is to borrow features from social networks.”
  • “Messaging apps like Whatsapp and Facebook Messenger are the norm when the young people communicate with each other…These SMS-like interactions (in the LMS) provide for almost realtime and instantaneous feedback that is perhaps even more convenient than face-to-face meetings.”
  • “We have online video clips in place of lectures and we even have a specialized forum to allow students to ask questions that are tagged to a specific point on the video timeline. This allows the teaching staff to understand the context of the questions and helps us identify the parts of a recorded video that might be wanting or not sufficiently clear.”
  • Gamification, together with timely feedback, provides the students will a sense of progress and this helps to improve student motivation and engagement…….I have also emphasized that gamification is not some silver bullet that will magically make teaching better. Consider it an icing that can make a good cake even better. If a cake is bad, no amount of icing can make it a good cake.

teaching is an inherently social activity, and so is learning which inherently has a social aspect to it. that’s why many teachers, during this time, choose to adopt ‘live’ meeting/lesson via Zoom/Google Meet/any other conferencing tools to engage students. so two questions that may lead on from here: (1) how do we ‘distribute’ the screen time limit among the different subjects each day? (what we are struggling here is screen time WITH teachers; some children are glued to screens big and small in a day and may very well exceed that ‘limit’?); (2) what other means/channels have we put in place to support the social phenomenon? looking at the 5 items that contributed to the successful implementation, [2] activity feeds, [3] in-platform messaging, and [4] in-platform forums are probably missing in popular platforms teachers are currently using (incl. Google Classrooms).

students need to be taught how to be online students. these would include the various aspects of a student’s life, including but not limited to attending lesson, doing and handing up assignments, clarifying doubts with teachers, receiving feedback and seek further clarification for understanding. and some of the rules and routines i have improvised so far include:

  • attending ‘live’ lesson:
    • join lesson with real names known to the teacher
    • use the teacher greeting students routine during f2f lessons to signify the start of the lesson (improvise and adapt the 行礼 bowing part)
    • mute mics if they introduce disturbance to the air (or teacher can choose to mute all mics when the teacher is talking)
    • whenever a question is initiated, a response is expected and students can feel free to unmute themselves to answer
    • use the ‘raise’ hand /like function in Zoom to indicate completion of tasks, or simple acknowledgement to teacher’s query (hmm… how to raise hand in Meet?)
  • doing and handing up assignments:
    • where paper-n-pen work is essential, use Adobe Scan to take photos of written work, and submit it via assignment set in Google Classroom
  • clarifying doubts
    • instant messaging via WhatsApp with photos to support teacher’s understanding of queries as necessary
  • receiving feedback and further clarifying
    • read the comments provided by teacher, and WA as necessary (ownership of learning)
    • teacher may proactively check on students’ understanding of feedback via WA too

i believe the list will expand and be refined as our experiences grow. oh btw, the above list applies to my secondary school students. if you are a K-6 primary school teacher, your list will likely to be different (: now, it’s time for me to go back to my lesson designing and preparation … …

#SGUnited #TogetherWeWillImprove #TogetherWeWillOvercome

(p.s. prof leong has published a note on fb yesterday to share his thoughts on the recent Zoom bombing event and the ongoing hbl)

记iMTL 进阶课 in-progress

yesterday was the second f2f lesson of our iMTL 进阶课 (wondering why the english course title uses ‘intermediate’?). the first f2f was exactly 1 month ago. in this course, the (old) title (which 露丝 and I din bother to change) included keywords ‘SDL’ and ‘CoL’. so, i had used the first hour of yesterday’s session to conduct a co-generative dialogue on 自主学习 (self-directed learning). and the following is the reification of our discourse:

14/8 discourse artefact – 自主学习 SDL;participants: 馨云、一凤、伟如、美娴、崇蕾、侯琳、玉侠、春容

towards the end of the conversation, i was pleasantly surprised that most, if not all, of the key ideas of SDL were surfaced without me needing to introduce them. sounds like a phenomenographic dialogue (if such thing exists; if not can cite Tan, 2019). realised this when i ran through the deck of slides adapted from a 12470 Clinic conducted some three years ago. was interested to see what the individually meaningful ideas that the teachers (馨云、一凤、伟如、美娴、崇蕾、侯琳、玉侠、春容) had created, i allowed them to complete a 2-qn survey via google form:

reflecting away, with qr code on the left, and the discourse artefact on the right in front

and here’s what they wrote:

q1. 我原来以为 … …q2. 我现在认识到 … …
让学生自己完成作业就能算是自主学习。自主学习包含了教师充足的设计与准备,让学生能够明白自己需要学什么、提供选择并给予学生自我评估的机会。
促进小学生完全自主学习是可能的,现在明白这个能力需要时间的发展。在小学阶段培养学生自主学习的能力是十分重要的。
Self directed learning 很容易达成学生需要确定学习目标和监督学习进展等
SDL就是E learning SDL 可以和很多东西有关。比如:老师的角色是什么、学生的学习效率如何
学生只要是能独立完成学习任务就是自主学习(如ezhishi)自主学习也有分几个等级,可以让学生主动找出自己需要学习什么,并自主地去提问,找答案。
自主学习是让学生单独、独立地学习在小学阶段,老师还是得先设计出学习活动让学生参与。适当地使用ICT能提高教学效应。
自主学习是学生有不同学习方式的选择。我们现在正处于学生是被动状态,要想让学生养成自主学习的习惯需要长时间的培养。
自主学习就是要让学生学习时有自主权,依据学生所需要或缺乏的那一块,提供更多的学习活动,辅助他们,让他们能够成功学习。的确如此。接触了一个新的概念solo taxonomy
自主学习只是做活动对主动学习有更深一层的理解,例如:学习动机、让学生能有想要学习的动机以及引学练用评和自主学习有密切的关系。

looks like 满满的学习 to you? and it’s not through transmission (of knowledge).

looking forward to the next session on 14 Sep for more knowledge creation (:

mobile first, a musing

a recent buzz word related to development at work — mobile first, or more specifically mobile first design — is today a key consideration for content providers and technological solution builders as mobile devices are everywhere, and these devices serve as one key source through which we consume contents around us. but, is mobile first the sole consideration, where learning is concerned? well, how about this — if there’s a first, there’s a second? mobile first certainly does not mean mobile #only; so what’s next?

(acknowledgement: (c)联合早报 副刊 Lianhe Zaobao 5th Jun 2019)

any discussion of technology and learning (or ‘digital learning’ some would prefer), it is always #ThinkAffordances — what is your design/objectives of learning, and how can technology (e.g. feature/function/tool/app/platform) enable your desired learning design? those of us who have experienced it would agree with me that chat (aka instant msging; e.g. WhatsApp), and even fb wall post+comments do not quite afford deep discussions that lead to rise-above of ideas. expressing individual’s ideas is easy; but to build on each other’s ideas, extend the ideas, improve on the ideas, counter-propose arguments, provide different perspectives, many of these acts of co-creating a common meaning is lost within the scrolling messages. in short, fb wall post+comments set up do not afford deep (digital) learning, currently.

so, what does? discussion forums (DF) do afford deep learning as described above (of cos, i will propose that knowledge forum (KF) is specially designed for this and is thus good-er). however, DF as we know it traditionally (oh btw, it’s a 20+ yr old technology), does not go well with the idea of mobile first design. the way discussions’ organised on screen traditionally, including threading, quoting, branching off, referencing back to other posts, is difficult to automatically fit on screens while displaying the discussions in way visually good for participants or readers of discussions. DF, in order to become mobile-first needs a rethink of its good old interface.

the term DF still invokes a tool-centric perspective. taking the affordance perspective, we can, and should examine what are the affordances of DF (and even KF) in enabling deep learning, and create them within existing mobile-first friendly solutions. in the process, DF as we have experienced it, may become a past, and it’s perfectly fine from a technological advancement perspective. the affordances of DF lives on.

going back to the fb wall post+comments idea, how can we improve it to afford deep learning per DF? first, we will need to invoke the concept of sub-walls (cf. discussion threads; sub-forums), where there can be more than one walls (cf. fb). and here’s probably a list of ‘requirements’ i would like to see:

  • there can be any number of sub-walls as a learner chooses to maintain.
  • to create a new sub-wall, one can select any existing post from the main wall or any sub-wall (together with existing likes, sharing info and comments), and pipe it into the new sub-wall.
  • each sub-wall has an unique identifier (e.g. a URL) that enables one to link back or to point to it in any post, comment in any (sub-)wall.
  • sub-walls will be arranged in reverse chronological order by default; and so are the posts and comments within.
  • one are allowed to pin any sub-wall, or any post within it, so that it is always on-top.
  • one can create and add a title/topic for each sub-wall to indicate the discourse topic to other participant/viewer.
  • one can define and add tag(s) to a sub-wall to facilitate personal classification, search and retrieval
  • everyone can add comments to any contents within a sub-wall
  • when adding a comment in sub-wall, one can upload images, documents, media, hyperlinks as part of the comment
  • visually, one should be able to easily identify any sub-wall that contains a new activity or reply/ to him/her; the ‘new’-ness visual cue is individualised and user-specific.
  • visually, within a sub-wall, one should see up to 3 levels of hierarchical nesting for comments to a post. this hierarchical view allows one to easily identify the relationship among comments and to focus their attention over selected comment(s) as desired.
  • participants will receive notification of new comment added to a sub-wall to enable them to return to continue the discourse (viewer can choose to follow and be notified too)

to end off, #mobilefirst (design) is not #mobileOnly (design); learning sciences folks can do wonders with computing folks when the two groups pool their brains together (:

to end off x2, finally, may i wish my Muslim friends:

(acknowledgement: creator unknown; as seen shared by darren)

[afternote] a note for self, it took 5 hours for this post to be first mooted in the morning while flipping newspaper, to connecting it with an idea@work, to typing and organising the ideas, to finally publishing it to the www & fb. now, who says learning (online) is fast? not for the slow-learner-me definitely (: