Adrian Zmudzinski (@CriptoAZ) header
Adrian Zmudzinski (@CriptoAZ) profile picture

Cryptocurrency journalist at @Cointelegraph Former @decryptmedia, @Benzinga Ideas are my own

Turin, Piedmont
Joined February 2018
Photo by Adrian Zmudzinski (@CriptoAZ)
My real Telegram profile is @CriptoAZ, @AdrianZmudzinski is an impersonator, be aware. I do not promote any investments, I do not give away or ask for crypto.
16
17
Adrian Zmudzinski retweeted
Hahahahahaha VPNs are HURTING CHILDREN Hahahaha fucking stupid fucks
Virtual private networks #VPN are increasingly used to bypass online age verification. Protecting children online is a priority, with new rules being implemented requiring a minimum age for access to some services Read👉 link.europa.eu/FGfr6C #DSA @EP_Justice @FZarzalejos
Readers added context they thought people might want to know
According to the University of Michigan research paper "Multi-perspective study of VPN users and VPN providers," 82.1% used VPNs "to protect myself from various threats/adversaries." No research shows that VPNs would be increasingly used to bypass online age verification. censoredplanet.org/papers/VPN-Sur…
147
2,174
24
21,249
526,126
.@nero_eth Hi, I'd like to ask you a few questions for a potential @Cointelegraph article. PM me if possible!
1
141
.@v_a_n_t_a_n_a Hi, I'd like to talk to you for a potential @Cointelegraph article. I think you really want to talk about what I have questions about. PM me if possible.
118
Adrian Zmudzinski retweeted
Your $200 in this market is $2M by Monday
264
50
22
601
133,640
Adrian Zmudzinski retweeted
Terrorismo, nasce il reato di autodocumentazione per i bombaroli fai-da-te. Chi cerca informazioni su internet rischia da 2 a 6 anni dlvr.it/SysxRR
154
15
61
31
101,802
INDI GREGORY, PERCHE' NON E' CURABILE (thread) Due righe su Indi Gregory. Quando si parla di bambini è sempre un disastro comunicativo perchè si rischia di essere eccessivamente retorici o cinici, e in definitiva è sempre difficile esprimersi su questioni spinose.
62
196
47
904
107,288
Adrian Zmudzinski retweeted
Non so se esiste un limite.
1,241
3,426
409
19,820
907,867
Adrian Zmudzinski retweeted
Is the title "AI Prompt Engineer" an upgrade or downgrade from "Chief Metaverse Officer"? Asking for a friend.
24
7
1
76
9,753
Adrian Zmudzinski retweeted
You have two options: Tile the forested mountains with black silicon panels and get 50-100 megawatts for 25-30 years. Put a BWRX-300 in an IKEA-sized building and get 290 megawatts for 60-80 years. Which is better for the environment?
672
2,014
301
11,995
4,072,627
Compelled to try and understand why the following is so difficult… - Believe that Isreal deserves the right to live in peace while empathising with the 2m Palestinians living under occupation - Sympathise with ordinary Palestinians while condemning Hamas in the strongest possible way - Condemn the celebration of Hamas attacks while understanding why Palestinians are so angry - Be disgusted by Hamas and their campaign of murder and kidnap while fearing for the innocent in Gaza when the full force of the IDF hits - Believe that Isreal deserves the right to defend itself but recognise the IDF has also committed human rights abuses itself Sympathy for ordinary Palestinians who live under both an occupation and murderous dictatorship does not make you an antisemite or terrorist sympathiser. It is sympathy for people who live in poverty under perpetual war. I am shocked and disgusted by what Hamas did yesterday but I find the calls to “bomb the fuck out of” and “glass Gaza” difficult to read. We’ve seen this cycle for decades now, it’s just more death, more destruction, deeper division. All people deserve dignity and the right to live, work and support their family in peace. It is not strong men who drive conflict, it is the weakest men.
251
178
31
1,841
353,821
Adrian Zmudzinski retweeted
Me ignoring the cookies from a site while still navigating
1
2
1
28
1,345
Use ChatGPT plugins without Plus subscription Using @langchain You can execute ChatGPT plugins for free With ChatGPT api In < 10 lines of code Details below ↓
12
54
1
267
90,732
Adrian Zmudzinski retweeted
Never change, Poland
523
3,968
584
66,252
4,522,809
Adrian Zmudzinski retweeted
Proprio oggi, #25aprile, annuncio la mia battaglia e dura resistenza contro una fetta della mia categoria/corporazione. Lotto contro quel lembo marcio e fascista, egoista e NoPos, diffamatorio e razzista che mi circonda, che mi attacca e mi isola sempre di più. Bona lè. ⤵️
211
431
80
4,423
384,403
Adrian Zmudzinski retweeted
LlamaAcademy: a factory that teaches LLaMA's how write API code. ChatGPT Plugins, but no manifest file is needed -- it just reads the API docs like a human would! For example, here is a finetuned model that learned how to talk to Notion by reading its API docs:
19
109
23
878
582,597
Adrian Zmudzinski retweeted
There's a chance that LoRA finetunes work so well that it dramatically alters the finetuning vs. retrieval + few-shot prompting power dynamic in favor of the former for many applications. PEFT (Parameter Efficient Finetuning, LoRA included) are emerging techniques that make it very cheap to finetune LLMs because most of the parameters can be kept frozen and in very low precision during training. The cost of pretraining and finetuning decouple. huggingface.co/blog/peft +LoRA (the code is very short/readable) github.com/microsoft/LoRA
LlamaAcademy: a factory that teaches LLaMA's how write API code. ChatGPT Plugins, but no manifest file is needed -- it just reads the API docs like a human would! For example, here is a finetuned model that learned how to talk to Notion by reading its API docs:
34
231
27
1,513
408,159
Adrian Zmudzinski retweeted
HOLY MOLY they acknowledged it! > from the US Treasury's first risk assessment on DeFi:
59
479
72
2,098
256,514
Adrian Zmudzinski retweeted
I've developed a lot of plugin systems, and the OpenAI ChatGPT plugin interface might be the damn craziest and most impressive approach I've ever seen in computing in my entire life.
80
679
78
5,718
1,619,096