Hello Everyone, The original post about this project can be found [here.](https://peakd.com/hive-169321/@jacobpeacock/building-and-testing-a-hive-custom-gpt-model-for-on-boarding-development-and-informational-purposes) The Hive GPT model can be found here: https://chat.openai.com/g/g-rf1eXIlTZ-hive Please note that you will need a ChatGPT Plus subscription to use it. # Some notes on Hive GPT one week in! Having already launched and to some degree used the custom GPT model tailored for Hive for roughly a week now I thought to do a followup article. I do not plan to get extremely detailed or too 'in the weeds' so to speak... but I do want to convey some ideas and insights that I have had along the way... and perhaps a path forward for the project where it will not have to rely on the ChatGPT servers. To be clear... the Hive GPT model that I am referring to is still a 'proof of concept' and I think that other folks should be building them also... or at least weighing the options and investigating the technologies being applied in the current model... so they can be built. Ideally having such a model that operates in a decentralized manner via the Hive blockchain itself would be the best long term goal... but for now (for my part) I would like to see such customized models being made available to as many people as could use them. Simply put I think they can solve a big part of educating folks about Hive in one of the most painful parts of the Hive learning curve... the initial technological hurdles mingled with grasping (even in the barest way) what is 'going on under the hood' with it all. It should go without saying that I dislike the current model being behind a 'Plus' subscription on ChatGPT but it truly was the only option that I was aware of at the time... and honestly in the end (with the current state of affairs and available services) it might be the best place for it. All of which to me means that creating more models (either near or of equal quality) is all the more important... and especially ones not hosted on centralized servers. Another thing worth exploring is models that are tailored towards individual aspects of Hive and its ecosystem. The current model is a generalist model with a heavy emphasis on education and development but a model could be created explicitly for creating content or doing curation. To be clear there I mean a model that helps guide a new user through the nuances of various frontend interfaces as well as proper tagging, utilizing markup and so forth. As a side note to all of that... I am in no way, shape or form picturing the custom models ever being used to create content aside from *maybe* illustrative technical diagrams and such for STEM related content. I think that should be quite apparent for my part... and a point that I hope others will follow up on... if they also choose to create their own custom Hive GPT models. I feel like I cannot stress that enough here and I hope others see that it is a tool for onboarding, education, information and development to be used for the betterment of the Hive ecosystem as a whole. It is also worth noting that the more that folks either use the existing model (or use ones that others create) the more that the model will learn and fine tune its own knowledge base. In the case of the current model being on the service it is on... everything it learns has the potential to *then* be learned by all their other models via the training data mechanism. I am not saying that last bit is good or bad... but it is what happens. When ChatGPT first came on the scene it knew pretty much 'squat' about the Hive blockchain (or even Web3 in general) and much of what it knew was not just mildly 'off' but either flat out bogus, a mix of 'near truths' or some other such conglomeration of misinformation. Not to get lost on a side tangent here but as of now all their public models are very well versed not just about the Hive blockchain but alas also its rich history... in a clear and concise manner. Do not get me wrong there because such models are still prone to errors... but my point being is that I witnessed its transformation upon the topic over time... and think that they will continue to do so. It is well worth noting that the current ones actually do even better at not making errors because their respective 'cut off date' for external data (the internet) got moved ahead a good bit so that they are more up to date. Recently I saw that [GPT-Baker](https://huggingface.co/spaces/abidlabs/GPT-Baker) had been released on Huggingface but it needs a 'HF Pro' account to operate it so I do not know all the 'ins and outs' of how it works or what all is truly possible with it. I did notice that the feature for adding external files for the model to use via a code interpreter was not functional. So far it is the only option that I have seen out there but I am sure more will crop up over time. In other news I have also been listing (and getting listed) the current custom Hive GPT model on some sites that more or less operate under the premise of a 'GPT Store' for custom models. They are not for sale or anything and it is much more akin to an open list of GPT models made by folks... so I figured that it would not hurt to make some submissions. I have started looking for more such 'stores' but so far the others either want a fee or I need to create an account of some kind to use them. Places that I either listed the model or got it listed are... The OpenAI developer community here: https://community.openai.com/t/show-off-your-gpts-in-here-for-the-community-to-test-2023-gallery/491656/442 The 'Awesome-GPT-Store' here: https://github.com/Anil-matcha/Awesome-GPT-Store This is where the 'Awesome-GPT-Store' does the new list hosting: https://www.thesamur.ai/custom-gpt-store I really should ramp up my efforts on getting it listed in more places but for now I am still in a bit of a 'holding pattern' until I have more of an understanding about how useful the project is overall... and if there is even much demand for such a model. Given that it is currently stuck behind a 'ChatGPT Plus' membership I really dislike it being the only option because it excludes too many folks from having direct access to it... and limits how much feedback I am going to get about the model itself. Okay, one last thing worth noting here is that I am working on a system where I can possibly get a lot more of the various Hive code bases all uploaded to the current model but it will add some complexity to the scenario for the 'average' user so I am hesitant to deploy it. Currently (with the ten file limit) I think I included the most important aspects (code) to 'get someone started' developing on Hive... without mixing all the projects together by placing them in a single compressed archive or 'zip' file. As things now stand with the model... it works quite well and is not overly complicated to use which is why I want to reiterate the importance of making 'focused models' and letting the current one remain a 'generalist' that does a little bit of everything... which makes for a good 'introductory' model for new Hive users or those just curious about Hive. Okay, that is about it! I hope that everyone is doing well and has a nice time. <center></center> <center>*The Hive GPT image is so cool I am going to use it in these reports!* [*Source*](https://peakd.com/hive-169321/@jacobpeacock/building-and-testing-a-hive-custom-gpt-model-for-on-boarding-development-and-informational-purposes)</center> <center>**Want to join Hive?**</center><center>**Sign Up Via My Referral Below!**</center><center>https://peakd.com/register?ref=jacobpeacock</center> <center></center> <center>Thanks for reading!</center> ## <center>BTC Donations & Tips are appreciated!</center> <center>bc1q0hgsylf3e5g6ycd6mdv0cs6cffvh5epy7233xm</center>
author | jacobpeacock |
---|---|
permlink | hive-gpt-report-1 |
category | hive-139531 |
json_metadata | "{"app":"peakd/2023.11.1","description":"Notes on the custom Hive GPT model after one week and some thoughts going forward.","format":"markdown","image":["https://files.peakd.com/file/peakd-hive/jacobpeacock/AJoHKwrbu7N744JHR1B7wTju7SJCYKbpLWYKA9QLStYrjutVT7W5ajYmjkx8mBA.jpg","https://files.peakd.com/file/peakd-hive/jacobpeacock/23t7YbEZXP3PDrCTtbh7izqfvv5DNwivT4LsVScHmhc4m6EGYLCrjp63q9v6DQ8edC24G.png"],"tags":["hive","palnet","neoxian","vyb","coding","ai","development","programming"],"users":["jacobpeacock"]}" |
created | 2023-11-24 03:05:42 |
last_update | 2023-11-24 03:20:42 |
depth | 0 |
children | 7 |
last_payout | 2023-12-01 03:05:42 |
cashout_time | 1969-12-31 23:59:59 |
total_payout_value | 2.841 HBD |
curator_payout_value | 2.827 HBD |
pending_payout_value | 0.000 HBD |
promoted | 0.000 HBD |
body_length | 8,182 |
author_reputation | 324,212,435,774,352 |
root_title | "Hive GPT Report: 1" |
beneficiaries | [] |
max_accepted_payout | 1,000,000.000 HBD |
percent_hbd | 10,000 |
post_id | 129,121,811 |
net_rshares | 11,761,930,167,972 |
author_curate_reward | "" |
voter | weight | wgt% | rshares | pct | time |
---|---|---|---|---|---|
neoxian | 0 | 6,245,468,222,508 | 30% | ||
ausbitbank | 0 | 2,502,523,264,867 | 100% | ||
ancolie | 0 | 6,432,848,426 | 7% | ||
walterjay | 0 | 129,072,537,510 | 10% | ||
l337m45732 | 0 | 2,564,713,455 | 6% | ||
felt.buzz | 0 | 74,877,485,370 | 10% | ||
akomoajong | 0 | 890,183,994 | 30% | ||
elsurtidor | 0 | 90,451,384,474 | 100% | ||
twoitguys | 0 | 8,049,743,125 | 25% | ||
zonguin | 0 | 2,124,103,562 | 10% | ||
francosteemvotes | 0 | 2,915,067,978 | 20% | ||
evildido | 0 | 1,152,708,590 | 10% | ||
aamirijaz | 0 | 3,917,444,425 | 15% | ||
aidefr | 0 | 3,622,545,117 | 16% | ||
sorin.cristescu | 0 | 51,561,735,585 | 10% | ||
votehero | 0 | 53,680,966,552 | 11.4% | ||
mawit07 | 0 | 28,460,773,874 | 9.9% | ||
xabi | 0 | 777,985,807 | 20% | ||
dbddv01 | 0 | 1,447,787,492 | 10% | ||
joedukeg | 0 | 1,293,644,837 | 25% | ||
cahlen | 0 | 30,562,582,249 | 99% | ||
robotics101 | 0 | 6,143,331,080 | 20% | ||
lpv | 0 | 1,281,807,487 | 4% | ||
statorcrime | 0 | 653,590,897 | 10% | ||
spiritabsolute | 0 | 37,000,294,879 | 100% | ||
abmakko | 0 | 2,024,346,318 | 15% | ||
gribouille | 0 | 729,440,468 | 16% | ||
anikys3reasure | 0 | 657,970,805 | 15% | ||
movement19 | 0 | 544,535,802 | 2.5% | ||
p-props | 0 | 19,041,367,213 | 100% | ||
imcore | 0 | 864,303,010 | 10% | ||
xawi | 0 | 62,737,502,055 | 30% | ||
solominer | 0 | 913,228,520,882 | 10% | ||
ifeoluwa88 | 0 | 16,630,491,140 | 30% | ||
meher04 | 0 | 1,376,103,577 | 30% | ||
crypticat | 0 | 68,677,744,653 | 20% | ||
johndoer123 | 0 | 100,299,706,647 | 100% | ||
ajongcrypto | 0 | 867,393,517 | 30% | ||
revueh | 0 | 1,124,146,925 | 20% | ||
definethedollar | 0 | 926,925,120,955 | 50% | ||
sanjeev021 | 0 | 996,973,725 | 3% | ||
myjob | 0 | 36,479,436,755 | 100% | ||
wenchebakken | 0 | 41,612,144,953 | 50% | ||
akomoajong1 | 0 | 989,732,648 | 30% | ||
marriakjozhegp | 0 | 1,181,673,416 | 30% | ||
evans-education | 0 | 5,915,731,785 | 50% | ||
paradoxtma | 0 | 44,078,143,847 | 100% | ||
hivequebec | 0 | 1,383,285,887 | 20% | ||
laruche | 0 | 45,013,298,816 | 20% | ||
mami.sheh7 | 0 | 842,107,537 | 15% | ||
logicforce | 0 | 497,972,259 | 9.9% | ||
recoveryinc | 0 | 3,190,742,983 | 5% | ||
brofund-ag | 0 | 1,006,529,780 | 30% | ||
samrisso | 0 | 3,436,870,181 | 5% | ||
tomtothetom | 0 | 1,446,453,213 | 10% | ||
mythix | 0 | 37,918,196,286 | 50% | ||
efastromberg94 | 0 | 13,444,181,311 | 100% | ||
mythix.market | 0 | 33,102,603,408 | 50% | ||
merit.ahama | 0 | 20,061,642,744 | 9% | ||
memesupport | 0 | 503,808,886 | 30% | ||
nopasaran72 | 0 | 1,842,227,767 | 20% | ||
mythix.token | 0 | 20,676,174,523 | 50% | ||
cugel | 0 | 4,761,372,583 | 5% | ||
acantoni | 0 | 1,598,777,436 | 5% | ||
tydynrain | 0 | 5,338,161,931 | 10% | ||
waivio.curator | 0 | 1,705,880,981 | 3.17% | ||
taradraz1 | 0 | 581,765,261 | 100% | ||
dragonmk47 | 0 | 3,855,203,691 | 3% | ||
protokkol | 0 | 13,092,471,682 | 30% | ||
crypt0holics | 0 | 732,059,212 | 6% | ||
moremoney28 | 0 | 5,508,604,736 | 15% | ||
jhymi | 0 | 3,007,140,151 | 15% | ||
hellsveiah | 0 | 2,276,276,498 | 100% | ||
ijelady | 0 | 680,188,972 | 15% | ||
the-lead | 0 | 516,884,021 | 15% |
I was going to ask if you made the picture but then saw what you wrote under it. I have a friend named @tydynrain who has helped me and others with doing things on Hive and told him about this thing you made, and he liked it and said it is like an AI for Hive. I have not looked at it yet because it says I have to make an account to see it.
author | myjob |
---|---|
permlink | re-jacobpeacock-s4nipb |
category | hive-139531 |
json_metadata | {"tags":["hive-139531"],"app":"peakd/2023.11.2"} |
created | 2023-11-24 23:32:00 |
last_update | 2023-11-24 23:32:00 |
depth | 1 |
children | 3 |
last_payout | 2023-12-01 23:32:00 |
cashout_time | 1969-12-31 23:59:59 |
total_payout_value | 0.000 HBD |
curator_payout_value | 0.000 HBD |
pending_payout_value | 0.000 HBD |
promoted | 0.000 HBD |
body_length | 342 |
author_reputation | 175,052,777,157,423 |
root_title | "Hive GPT Report: 1" |
beneficiaries | [] |
max_accepted_payout | 1,000,000.000 HBD |
percent_hbd | 10,000 |
post_id | 129,143,928 |
net_rshares | 5,734,713,810 |
author_curate_reward | "" |
voter | weight | wgt% | rshares | pct | time |
---|---|---|---|---|---|
jacobpeacock | 0 | 3,085,411,477 | 100% | ||
tydynrain | 0 | 2,649,302,333 | 5% |
Yep, I guess calling it an 'AI for Hive' is not off the mark. Eventually I am hoping to develop (or see someone else develop) one that does not require a subscription account to use it.
author | jacobpeacock |
---|---|
permlink | re-myjob-s4nkwg |
category | hive-139531 |
json_metadata | {"tags":["hive-139531"],"app":"peakd/2023.11.2"} |
created | 2023-11-25 00:19:30 |
last_update | 2023-11-25 00:19:30 |
depth | 2 |
children | 2 |
last_payout | 2023-12-02 00:19:30 |
cashout_time | 1969-12-31 23:59:59 |
total_payout_value | 0.000 HBD |
curator_payout_value | 0.000 HBD |
pending_payout_value | 0.000 HBD |
promoted | 0.000 HBD |
body_length | 186 |
author_reputation | 324,212,435,774,352 |
root_title | "Hive GPT Report: 1" |
beneficiaries | [] |
max_accepted_payout | 1,000,000.000 HBD |
percent_hbd | 10,000 |
post_id | 129,144,842 |
net_rshares | 2,646,680,638 |
author_curate_reward | "" |
voter | weight | wgt% | rshares | pct | time |
---|---|---|---|---|---|
tydynrain | 0 | 2,646,680,638 | 5% |
That would be nice.
author | myjob |
---|---|
permlink | re-jacobpeacock-s4p6fp |
category | hive-139531 |
json_metadata | {"tags":["hive-139531"],"app":"peakd/2023.11.2"} |
created | 2023-11-25 21:02:15 |
last_update | 2023-11-25 21:02:15 |
depth | 3 |
children | 1 |
last_payout | 2023-12-02 21:02:15 |
cashout_time | 1969-12-31 23:59:59 |
total_payout_value | 0.000 HBD |
curator_payout_value | 0.000 HBD |
pending_payout_value | 0.000 HBD |
promoted | 0.000 HBD |
body_length | 19 |
author_reputation | 175,052,777,157,423 |
root_title | "Hive GPT Report: 1" |
beneficiaries | [] |
max_accepted_payout | 1,000,000.000 HBD |
percent_hbd | 10,000 |
post_id | 129,166,270 |
net_rshares | 0 |
Fascinating stuff! Thanks much for sharing! Using it as a means of educating and onboarding people sounds like a marvelous idea! 😁🙏💚✨🤙
author | tydynrain |
---|---|
permlink | re-jacobpeacock-20231124t221230232z |
category | hive-139531 |
json_metadata | {"tags":["hive","palnet","neoxian","vyb","coding","ai","development","programming"],"app":"ecency/3.0.37-vision","format":"markdown+html"} |
created | 2023-11-25 08:12:30 |
last_update | 2023-11-25 08:12:30 |
depth | 1 |
children | 2 |
last_payout | 2023-12-02 08:12:30 |
cashout_time | 1969-12-31 23:59:59 |
total_payout_value | 0.000 HBD |
curator_payout_value | 0.000 HBD |
pending_payout_value | 0.000 HBD |
promoted | 0.000 HBD |
body_length | 134 |
author_reputation | 206,721,900,138,544 |
root_title | "Hive GPT Report: 1" |
beneficiaries | [] |
max_accepted_payout | 1,000,000.000 HBD |
percent_hbd | 10,000 |
post_id | 129,151,790 |
net_rshares | 3,346,889,039 |
author_curate_reward | "" |
voter | weight | wgt% | rshares | pct | time |
---|---|---|---|---|---|
jacobpeacock | 0 | 3,346,889,039 | 100% |
Glad you like it! The educational potential is truly incredible. In many ways it is so much easier to learn stuff from an AI than it is from a human or even via self-teaching methods. 🤠
author | jacobpeacock |
---|---|
permlink | re-tydynrain-s4oqpe |
category | hive-139531 |
json_metadata | {"tags":["hive-139531"],"app":"peakd/2023.11.2"} |
created | 2023-11-25 15:22:30 |
last_update | 2023-11-25 15:22:30 |
depth | 2 |
children | 1 |
last_payout | 2023-12-02 15:22:30 |
cashout_time | 1969-12-31 23:59:59 |
total_payout_value | 0.000 HBD |
curator_payout_value | 0.000 HBD |
pending_payout_value | 0.000 HBD |
promoted | 0.000 HBD |
body_length | 185 |
author_reputation | 324,212,435,774,352 |
root_title | "Hive GPT Report: 1" |
beneficiaries | [] |
max_accepted_payout | 1,000,000.000 HBD |
percent_hbd | 10,000 |
post_id | 129,159,810 |
net_rshares | 2,132,634,830 |
author_curate_reward | "" |
voter | weight | wgt% | rshares | pct | time |
---|---|---|---|---|---|
tydynrain | 0 | 2,132,634,830 | 4% |
Indeed! Of all the uses for AI, this is the most interesting to me. I've never tried to learn for an AI, so I can't say, but it sounds potentially very useful! 😁 🙏 💚 ✨ 🤙
author | tydynrain |
---|---|
permlink | re-jacobpeacock-20231125t205355102z |
category | hive-139531 |
json_metadata | {"type":"comment","tags":["hive-139531"],"app":"ecency/3.0.44-mobile","format":"markdown+html"} |
created | 2023-11-26 06:53:57 |
last_update | 2023-11-26 06:53:57 |
depth | 3 |
children | 0 |
last_payout | 2023-12-03 06:53:57 |
cashout_time | 1969-12-31 23:59:59 |
total_payout_value | 0.000 HBD |
curator_payout_value | 0.000 HBD |
pending_payout_value | 0.000 HBD |
promoted | 0.000 HBD |
body_length | 170 |
author_reputation | 206,721,900,138,544 |
root_title | "Hive GPT Report: 1" |
beneficiaries | [] |
max_accepted_payout | 1,000,000.000 HBD |
percent_hbd | 10,000 |
post_id | 129,175,717 |
net_rshares | 2,816,002,559 |
author_curate_reward | "" |
voter | weight | wgt% | rshares | pct | time |
---|---|---|---|---|---|
jacobpeacock | 0 | 2,816,002,559 | 100% |