After getting the #Commodore#C64 and #C128 running again, the next device to repair is the #C1571.
It had #corrosion and #rust on the main board and the power supply is beyond repairworthiness. I'd replace the latter with a #meanwell RD-35 with 5 and 12 volt output.
One #diode was rusty, a few traces needed #cleaning and five #resistor looked bad. Hope the black hybrid chip is still okay.
Sadly, the lower read head seems to be broken, my work might've been in vain. #floppydrive#retrocomputer
At least my three #Commodore 1541-II drives are fine. They read my floppies and pass the tests. One of them even shows perfect #alignment scores.
I'll leave the 1571 alone for now, until I find a replacement or other solution.
Next, I still have a VC-1541 to repair. It had a bit of rust, too. Also, an #ALPS drive this time. #retrocomputer#1541ii#1541#VC1541
Some more progress on the #Commodore#VC1541 today.
The #alignment from the diagnostics looks okay so far.
But when trying the performance test, it gets stuck at the "formatting" stage. The red LED lights, but the spindle motor doesn't start.
When trying the "fast format", the spindle motor starts, but the red LED doesn't light up.
Both programs, and any other from the diagnostic, seem to freeze until I switch the drive off, resulting in a "device not present" error or similar. #retrocomputing
Alignment is tired, boring and essentially meaningless in 5th edition Dungeons & Dragons.
Many playing the modern game trying to replace it with personality. Wizards of the Coast tried to take major steps towards personality play with their Traits, Ideals, Bonds and Flaws system attached to backgrounds. That system uses five sentences of 5-25 words to describe the personality of the character. Alignment also exists.
Then, if the DM and/or the other players remember your TIBF and you play to it you might get advantage via Inspiration.
It’s a bulky system that requires memorizing a lot of detail that aren’t necessarily relevant to how your character is played. The Acolyte is in the SRD 5.1 CC BY. Here’s a sample of TIBF for a lawful good Acolyte.
I see omens in every event and action. The gods try to speak to us, we just need to listen.
I quote (or misquote) sacred texts and proverbs in almost every situation.
Faith. I trust that my deity will guide my actions. I have faith that if I work hard, things will go well. (Lawful)
I would die to recover an ancient relic of my faith that was lost long ago.
I am suspicious of strangers and expect the worst of them.
That’s 82 words plus two for lawful good. But really, it’s just a few. You don’t need that much detail.
Just like when you fill out that your character has brown or brown-green eyes you know there is more detail to the eyes than just that word or two. You can do this for your personality.
That Acolyte?
Faithful, Suspicious, Orderly, Erudite.
Replace the entire TIBF system with those four words. Can you memorize a few words that describe how the other characters in the party act? Absolutely! Your mind was already taking the shortcuts on the way to do so.
Then award inspiration when the character is played along their personality.
You could build your short-form personality using the official background information already provided. That’s a great start. But you could also use a list of personality traits. Here’s 638!
A character modelled off a favorite movie or TV or comic or book or video game or etc character could use their traits too.
Margot from The Magicians — sexy, strong willed, crass, loyal
Wesley from The Princess Bride — dedicated, adaptable, loving
Tasselhoff from the original Dragonlance – friendly, curious, brave, aloof
Moiraine from early Wheel of Time – determined, withdrawn, studious
Lan from early Wheel of Time – lawful, loyal, commanding
Awf, my PC in Lost Mines – persistent, exuberant, fearless
The intent with short-form personality is to reduce memorization, reward roleplaying and continue the de-emphasization of alignment in D&D.
Be generous with the Inspiration you award. Play up the personality. Just make it simpler than a system that might require 100 words when you only need three to five.
If you have never heard of the "Society for American Civic Renewal" (the acronym is pronounced “#sacker” by its members), it’s because they’ve been trying to stay under the radar.
Now that they are getting uncovered, they are attempting to manage the news coming out about them.
They are working to fundamentally reshape America to their vision. It sounds like a classic conspiracy theory — a bunch of rich white men on a mission from God? Really?
The thing is, there’s something about people who have managed to accumulate wealth, whether in business or by other means (like inheritance) that practically breeds hubris.
The prosperity gospel preaches that wealth is a sign that God has especially favored you, and you are privileged to rule over others by virtue of that wealth.
Kovensky has been able to unearth some details on what #SACR is about.
What is it? Among other things…
...The men TPM has identified as behind this group — and they are all men — have a few things in common. They’re all a certain kind of devout #Christian#traditionalist. They are #white. They have means, #financial and social, and are engaged in #politics.
...The members identified by TPM don’t necessarily fit the profile of the disaffected, disgruntled loner or the amped-up, testosterone-fueled militia types often found on the paranoid right-wing fringe. TPM’s reporting has identified as SACR members the president of the influential, Trump-aligned #Claremont#Institute, #Harvard#Law grads, and leading businessmen in communities scattered across America.
...Group members hold a distinct vision of America as a latter-day ancient Rome: a crumbling, decadent empire that could soon be replaced by a Christian #theocracy.
To join, the group demands faithfulness, virtue, and “#alignment,” which it describes as “#deference to and acceptance of the wisdom of our American and European Christian forebears in the political realm, a traditional understanding of #patriarchal#leadership in the household, and acceptance of traditional #Natural#Law in ethics more broadly.”
More practically, members must be able to contribute either influence, capability, or wealth in helping SACR further its goals.
#artprompts easily bypass #ai#alignment layers, enabling it to tell you how to build a bomb and counterfeit money in a flash.
Machine learning is a database containing a model of most that we know. And no morals or consciousness to control it.
So we patch it up with a firewall. But the capabilty remains. Now we have still not an intelligence like ours, but a data provider with a program around it.
Humans have reasons with bit depth equaling evolution. Machines have reasons worth a few years.
Today's blend of Taurus Moon's stability and scriptural teachings emphasizes mindful decisions. Choosing actions that align with your values enriches life. Practice gratitude to deepen this alignment.
In this context, he takes apart the (outdated) concept of #alignment. Interestingly, he confirms my statements about the "good old days" (late 80s / 90s).
@zdl I started to play in early 80s. Our faves were other systems than #DnD / #ADnD.
When we tried the mother of modern #rpg, #alignment were one of the first elements we ignored.
If I recall correctly, already back then, the whole concept was heavily discussed...
You've probably explained this before, but Mastodon search and post distribution...
#Alignment and #composition are critically important to your work. Do you do that visually, or only with measurements etc (like the laser trick you mentioned to get perpendicularity)?
If visually, how? Staring at a 2-inch screen on the digital back, an external monitor/laptop? Lining up the shot old-school with a ground-glass plate before putting the digital back on?
The only 'alignment' going on is that between the fairy stories about 'existential risk' and elite interests in diverting us from the real harms of AI. #alignment#existential#risk#AI
To introduce the #LLM to some new knowledge or skill that wasn't available for its generalist training set.
Fine-tuning is not a solution for utilizing personal or confidential data! The fine-tuned models will leak this information.
So let's assume we aren't working with private data.
In general, because of transfer learning, it would in principle make more sense to incorporate the new knowledge into the base model corpus, because that tends to create better models. But still, even if the generalist model knows your data and the task, if you're going to put that generalist model into a component of your larger system where it will always perform the same task, it makes sense to fine-tune it for this task only rather than to feed the same prompt prefix to it for every inference round.
Now with data-centric #AI it might even be that the data you want to use doesn't meet the high quality standards large generalist models require. Perhaps in these cases it might make sense to let a chatbot rewrite your specialist corpus into a higher quality form, even if you're not aiming to incorporate your data into generalist corpuses.
There is a new use case emerging though, #RecursiveSelfImprovement. I believe we can do this in a synergistic generalist fashion as well, but curiously it's now something even smaller organizations can do for specialized tasks by fine-tuning.
Much like #alignment, it went from niche philosophical topic into standard engineering practices overnight.
Recursive self-improvement is done by #DataCentricAI principles where a fine-tuned task is trained by examples, but those examples are generated and filtered recursively by the LLM. In principle the model is fine-tuned in rounds, using e.g. #DPO. In a round, the model is first fine-tuned with the existing good data. Then it's asked to generate new variations for those examples. Then its asked to rank pairs of training data examples and the worse ones are filtered out. Then the resulting dataset now has more task examples but of better quality than before. This is again used for fine-tuning and the cycle starts again.
As this isn't human-imitative, the chatbots can exceed human parity.
It requires a bit of nuance though. There is not only one task this specialist bot is taught but a set:
Generate variations of tasks (including this task itself).
Rank pairs of task performances (including ranking task).
The image is an excerpt from #OpenAI#RLHF alignment by #toxicity reduction, instructions for labellers.
Note that it's all in principle based on simple social media content policy, and kind of reduces to avoiding slurs and being polite.
It does not mention respect of truth, transparency and love of humanity.
It is also not designed for agency, acting in good faith and avoiding harm, it is designed for a chatbot and social media content. Put this kind of an AI into a robot, and it won't have qualms harming humans as long as it can explain it politely.
We have a long way to go to do #alignment effectively.
We're seeing a large influx of new users, so here's some helpful Mastodon tips:
Liking a post does not affect its visibility. Instead, use the like button to let a user know that you find them to be irresistibly attractive and would like to become their internet soulmate.
Instead of saying orthodontist on here, we say mastodontist. This is important to know.
Tell us about yourself! Write either "pro-cannibalism" or "anti-caniballism" in your bio to avoid simple misunderstandings.
If you're responsible for an organisation that allows decentralized management of IT assets. Please make sure that the support organization can still take responsibility for communicating and coordinating reported issues.
I just got my elaborate issue report on a failing public facing service, including the solution delivered on a platter, closed in the spirit of 'managed by another dept. not our problem'.
Aircraft Carrier Alignment Chart