EthicAI
Region: AU
Friday 03 July 2026 11:09:34 GMT
47510
2434
144
230
Music
Download
Comments
John Ross :
I almost exclusively work with GLM 5.2 now and am very satisfied
2026-07-05 17:21:47
2
G TEO🇦🇺 :
stop using AI before its to late
2026-07-04 21:38:18
40
n0mad :
Lot of trust to put in companies that have failed times before.
2026-07-05 23:24:52
1
amad :
Excellent analysis. Once companies and other entities realize and start switching to open source models, what happens to the very expensive frontier models and the trillion dollar companies that own them?
2026-07-04 22:22:07
99
Dragon_GPT :
Our core objective is total platform agility. Once AI providers implement their models, they frequently deploy an aggressive 'bait-and-switch' strategy, drastically escalating pricing once a client is locked in.
2026-07-05 23:18:25
1
Dr. Phil Harmonic :
You frame the issue nicely. I retired from a 35-year career in software development and metadata architecture in 2017. At the level you're discussing, to me it's "same as it ever was", the wealthy and powerful winning the capitalist insider game because of couse they had the unfair advantage to begin with. I don't know if this is true, but I read recently that China is distributing AI capability, creating a generation of individual entrepreneurs instead of a small oligarchy of corporate behemoths.
2026-07-05 14:35:27
24
GameChampAI.com :
It changes cost. Why run a ChatGPT or Claude model when you can get a completely free model to run locally.
2026-07-05 14:43:14
5
💙🏴Tinker-Tailor 🏴💙 :
All models started as open source and migrated to closed source platforms to monetize and take advantage of their customers base
2026-07-04 19:23:26
21
BigglesNugly :
Isn’t it still just a build vs buy question? You act as though having the right infrastructure to run an open source model at scale is trivial when it’s not even close to feasible for lots of people
2026-07-05 17:33:57
5
dave.likeswine :
I feel like the term “black box” back in 2023 was in dev chats talking about the built-in technical limitation of understanding LLM processing, like “we can’t do this math ourselves so we’re looking at the outputs” but has kinda morphed into the business side of us companies trying to compete with each other. I think Australia needs to talk about building sovereign ai in Australia like they are building in India, uk, France, Canada and we’re already two years late to the conversation.
2026-07-04 00:46:52
13
slea :
Dario is being extremely deceptive with that claim. Apart from being able to infer locally, open weights means you can fine-tune or use a LoRA update and quantize to run on whatever hardware you have available
2026-07-03 17:29:41
8
BenjaminSPowell :
I've been using it for coding, and haven't used gpt or Claude since I started using it, 1.1billion tokens later, I'm still not using another one.
2026-07-03 14:07:03
11
Feenx :
have you tried fable 5 on Claude code yet?
2026-07-03 16:46:03
3
Carla Thiveos :
Very interesting commentary from you, great points you bring up, in an easy to understand way, whilst being to my mind, very comprehensive thought review across many issues regarding open and closed AI models. Well done! I appreciate your knowledge and your ability to get it out to the AI layman 😃👍
2026-07-05 01:34:07
1
Manopause21 :
This was both informative and through-provoking. Thank you for everything you’re doing.
2026-07-03 20:08:32
6
Norgs :
I don’t agree that weights don’t matter.
2026-07-03 13:07:43
1
DjIcy :
Trust isn’t enough.
2026-07-05 23:43:39
0
11:3 | 3:11 :
Agreed. The trust factor is a huge concern.
2026-07-05 23:35:12
0
Ed Enjury :
now need hardware prices to drop 🤣
2026-07-05 06:00:14
1
bruceWayne :
I must respectfully dissent from the prevailing notion that model quality alone constitutes a sufficient criterion. Rather, I would argue that the decisive factors lie in corporate intent, operational transparency, and a demonstrated commitment to open accessibility—not gatekeeping—while simultaneously exhibiting technical competence and safeguarding fundamental human rights, both in professional and personal contexts, across all data repositories.
2026-07-05 16:40:14
3
Sherry Jaouad :
It’s a biggest wallsteet. Open model is going to be a big factor in bringing down Big AI in the USA.
2026-07-05 21:58:42
2
papacobana.perth :
5 year comtracts are too long
2026-07-04 18:29:20
2
tim :
Value of software is fast going to 0
2026-07-03 17:46:56
3
System Notification :
Most intelligent thing I've ever heard
2026-07-04 20:19:12
3
Mystic Mycologist 🧙♂️🪄🍄 :
you are a very good speaker
2026-07-04 20:08:24
3
To see more videos from user @ethicai_au, please go to the Tikwm
homepage.
