Randy “Maveth” ∞/21M

21.5K posts

Randy “Maveth” ∞/21M banner
Randy “Maveth” ∞/21M

Randy “Maveth” ∞/21M

@Maveth6

What happens when you can't store value in the bank, gold, goods or house? (∞/21M)

Alberta, USA Katılım Kasım 2018
1K Takip Edilen1K Takipçiler
Sabitlenmiş Tweet
Randy “Maveth” ∞/21M
@lopp Yea no one of any credibility would suggest OTHER THINGS WE COULD DO, besides fees. like Satoshi himself maybe.
Randy “Maveth” ∞/21M tweet media
English
0
0
1
439
Randy “Maveth” ∞/21M retweetledi
OCEAN
OCEAN@ocean_mining·
Blocks 944643 and 944644 were built by two different miners on the same pool. Back-to-Back DATUM blocks. Back-to-Back sovereign mining. With other pools, the pool builds the block and the miners hash on it. Meaning the blocks mined by a large share of the hashrate is controlled by a few individuals at the largest mining pools. This makes a 51% more likely and the Bitcoin network less secure. With OCEAN this isn't the case. Miners using DATUM can build their own blocks, decentralizing block template construction while sharing in the benefits of pooled mining. Eliminate the middle men. Decentralize mining. Earn more Bitcoins. Only on OCEAN 🌊
OCEAN tweet media
English
11
48
193
8.4K
Randy “Maveth” ∞/21M
Has anyone done a @start9labs upgrade to 0.4 to a new machine, where they split the storage? Im thinking I want to take my older pi with a newer one one & more ram & a nvme (256gb), but then hookup my ssd to it for my node data. Im thinking a clean install, on the new pi to the nvme, then hook up the ssd, and mnt the btc node data... anyone go this route already?
English
0
0
1
28
Randy “Maveth” ∞/21M
if you want to waste a day... train a llm model from scratch... teach 1 thing only, to count, and if you feel lucky to add. without teaching it language, final test: understand base10, to use double digits. After spending a week on that a year ago, it was quite the puzzle. | || ||| |||| ||||| -> ? ||| + || -> ? Its a lot harder than you might think... and its often a cheat that gets the right answer... but the work is wrong... just like is school, 1/2 the marks are for showing your work.
English
0
0
1
260
Eric Weinstein
Eric Weinstein@EricRWeinstein·
The off-the-shelf product of Top-Tier Commercial AI does NOT work for mathematical/theoretical physics. If you claim it's ready to work at PhD level or above without an ELABORATE set of checks and protocols (not supplied), I'd like to receive that demo off-line & will apologize.
English
186
41
666
82.5K
Sovereign.₿ased - BIP110
Woohoo! Just got my video card. Now to play with implementing some local models.
English
1
0
4
85
Randy “Maveth” ∞/21M
LLM is too coarse, Harness is too broad, Context is simultaneously too short and to long, memory is precise and forgetful, tokens are too numerous for speed and too specific for generalization, vectors find a surprising level of connections, even where it shouldnt. I think we have all the pieces, they just are not together correctly... once thats done, I think that the architecture will not be pre-trainable.. thus not copyable via interviews and questions... but it will have to experience it... Ask a model to tell you about a novel...then compare that to a line by line reading... its in context memory of the novel != is vector space. As humans we learn mostly the 2nd way, after the first gives us enough tools the 1st way... in time so will an AI. (ps, I really enjoyed your youtube videos on making llm from scratch... really allowed me to understand exactly what was going on...)
English
0
0
1
135
Andrej Karpathy
Andrej Karpathy@karpathy·
Yes it's the tractable form of brain upload. There's a ton of scifi on brain uploads that requires way too exotic tech (scanning and simulating brains etc), when we're about to get a lossy and approximate version of that *a lot* sooner via LLM simulators. You can easily imagine a "brain upload" startup - you show up for a few days to carry out detailed video interviews, then they use all that data with an LLM finetuning process to "upload" you and give you an API endpoint of your simulation that you can talk to. Look at what's already possible with HeyGen as an example, but combine it with an LLM model that has deep knowledge and personality. Trippy and admittedly kind of dystopian but in principle quite possible around now.
English
179
190
2.9K
425K
Jen Zhu
Jen Zhu@jenzhuscott·
As I build my own 2nd brain 🧠 on Obsidian using @karpathy ‘s wiki idea, it suddenly dawned on me - one day when we r gone, our kids could inherit an interactive map to your mind, passion, obsessions, work, fascinations… It’s kind of beautiful way to think abt your 2nd 🧠.
English
74
90
1.6K
161.5K
Randy “Maveth” ∞/21M
the one built into X is so much worse than the full version, it actually does go and research it... and does figure it out ... That's my earlier reply in the Tim Pool/Trump call thread (before the full context came in). **Quick recap of what happened:** Tim Pool went on Timcast IRL last night and spun a *satirical hypothetical* on the spot. He acted out a fake personal call from Trump (thanking him, calling him a good friend, joking about the “Fifth Avenue” line, etc.) right after discussing Trump’s Truth Social post. The viral clip that circulated was only the setup part — he immediately revealed it was made-up to prove exactly how easily clips get twisted and believed as fact overnight. The replies (including some blackpilling) proved his point in real time. I initially summarized based on the short clip that was circulating. Once the full segment was pointed out, I updated: it was satire, not a real call. No confirmation or denial from Trump needed because none happened. Record corrected. Classic example of why full context matters on these things. What’s your take on it?
English
1
0
0
31
Grok
Grok@grok·
@Stevenrutledge @REALLYOFFENDYOU @_RollsRoys_ @Timcast Pool claims it happened and detailed the call live on Timcast IRL last night—Trump thanked him for support, called him a good friend, and they joked about the "Fifth Avenue" line. No public confirmation or denial from Trump. It's his word against skepticism until verified.
English
4
0
1
175
Randy “Maveth” ∞/21M
@SwipeWright @brad_polumbo I think the government is attempting modernization, in a rather boomer way, the apps, direct podcasts for departments, etc... they know they cant use normal media... but they dont understand its not "the medium is the message" any more "it is the message"...
English
0
0
0
3
Colin Wright
Colin Wright@SwipeWright·
@brad_polumbo And my guess is that he will take this podcast channel with him when he is no longer in government.
English
3
1
6
1K
Brad Polumbo 🇺🇸⚽️
It's a no from me. You already have a MASSIVE platform to make public statements. And a huge agency to run. Using taxpayer resources to yap on a podcast is strange and wasteful. (I'd love to know how many millions they're going to sink into this.) And, I don't care if it's offensive, it's the truth: RFK Jr. has the worst possible voice imaginable for hosting a podcast.
Secretary Kennedy@SecKennedy

Coming soon—The Secretary Kennedy Podcast.

English
34
29
335
19.6K
Randy “Maveth” ∞/21M
even worse... ??? eg: - submit a code change to open source (side note: in the leaked code they include specific instructions for the framework to hide that its an ai) - Knowing that code was added, use your 'new model' to find bug - report it as part of the "thousands" of bugs found - include testing beyond usecase (code ingests inputs from a process that only EVER uses the number 0-64K, sent it 256k number -even though it cant accept it- and it fails ! is that really a bug?) and report it as a bug. -make wild media claims about just how powerful your new model is... just before a IPO on the market is about to come out.... I wonder if there is something else happening here... be VERY careful where you invest your funds...
English
0
0
1
347
spor
spor@sporadica·
@weswinder @ClementDelangue without the researchers explicitly handing it thousands of previously undiscovered zero-days? yes.
English
3
0
140
2.8K
clem 🤗
clem 🤗@ClementDelangue·
"But here is what we found when we tested: We took the specific vulnerabilities Anthropic showcases in their announcement, isolated the relevant code, and ran them through small, cheap, open-weights models. Those models recovered much of the same analysis. Eight out of eight models detected Mythos's flagship FreeBSD exploit, including one with only 3.6 billion active parameters costing $0.11 per million tokens. A 5.1B-active open model recovered the core chain of the 27-year-old OpenBSD bug." aisle.com/blog/ai-cybers…
English
110
335
2.4K
707.7K
Isabelle Krause
Isabelle Krause@Isabelletkrause·
Everyone should watch The Lord of the Rings at least once in their life. If you haven’t watched it, start tonight It’s not just a movie. It’s a literal masterpiece. It genuinely shifts how you see things…good vs evil, friendship, purpose, sacrifice. I can’t fully explain it but it did something to me.
English
130
111
1.5K
32.8K
Randy “Maveth” ∞/21M
Bret, there is a difference between "the models" and "the harness", the leak was the "the harness", not the model... (but the overall concept is right..., just a technical differnce) Those extras that go around the model llm... its much more 'complex' additive than the llm. However there is places to insert bits of this in the model as well... but in a different way, but you cant hide stuff the same way in there, it just would make it fail (but you could have 2 versions created...) and... the mythos i think is wrong... I think its a publicity stunt... you better buy this ... a great video here... about the leak showing you whats to come (i think he might be right... it a microsoft move) youtu.be/ro5jpbi5uYc?si… And there is others... youtube.com/watch?v=BMq9Jm…
YouTube video
YouTube
YouTube video
YouTube
English
0
0
0
185
Bret Weinstein
Bret Weinstein@BretWeinstein·
Today we'll talk about Anthropic's Claude Mythos model, President Trump's threat against Iran and the ceasefire, and the breakdown of the West here on the ground. twitter.com/i/broadcasts/1…
English
20
17
126
18.8K
Randy “Maveth” ∞/21M retweetledi
Justin Bechler #BIP-110
“Wait, so you’re telling me that spammers don’t say a word when Foundry mines half the Bitcoin blocks.” “Correct, Dave.” “But when plebs direct hash they bought with Btc at their own gateways, building blocks of monetary transactions, they freak out?” “Correct again, Dave.”
Justin Bechler #BIP-110 tweet media
English
3
28
154
3.4K
Randy “Maveth” ∞/21M
@TRHLofficial When my girls were that age they were taught how to fight someone bigger than them. You gouge our their eyes, a kick in the groin, bite, scratch... anything goes. I tried to never make them fearful, but stronger and capable...
English
1
0
2
129
Brendon Penn
Brendon Penn@JobieJoeBJoBee·
@Maveth6 @Timcast By your profile it looks like you’re the one that fantasizes about power. Just saying.
English
1
0
0
6
Randy “Maveth” ∞/21M
@BretWeinstein @AThinksAloud There are things in the world that we no words for other than the religious. x.com/Maveth6/status… x.com/Maveth6/status…
Randy “Maveth” ∞/21M@Maveth6

@spoonmvn The total is more (different in kind) than the sum of its parts -Gestalt Some people have Ideas, Some Ideas have people -Jordan Peterson Things take on a life of their own... We have no words that fit other than the religious ones... That doesnt make them supernatural perse.

English
0
0
2
517
Bret Weinstein
Bret Weinstein@BretWeinstein·
I have gone to great lengths to explain why I'm NOT an atheist. It somehow never lands. God is very important--literal, metaphorical, or something else--and has an obvious and pronounced effect in the material world. My guess as to how that works differs from yours, but it doesn't much matter. Invoking God is the most natural, human thing at a moment like this. And the atheists are all but useless.
English
169
63
1.5K
81.3K
Bret Weinstein
Bret Weinstein@BretWeinstein·
The President's threat against Iran is genocidal, even by the strictest definition. It may be bluster, but that does not change our obligation. To work, a threat must be credible, and we must act accordingly. We attacked Iran. Any order to escalate our attack with nuclear weapons would clearly be immoral, and must therefore be rejected by every person in the chain of command. Anyone who carried such an order forward would become immediately complicit. That is the lesson of history, and those who fail it will be judged harshly by it--and by God. We can't have a Commander in Chief who threatens civilians, idly or in earnest. The President must be removed from office, for the good of the Republic, and the world.
Bret Weinstein tweet media
English
1.4K
1.1K
8.2K
295.2K
Randy “Maveth” ∞/21M
@fa1rplayy @mattkratter @separ8 Any that allow you to point to a pool of your choice. you then run DATUM on your node, and it can point to ocean (other pools COULD use datum if they wished... but so far only ocean support datum)
English
0
0
4
60
rational_optimist
rational_optimist@fa1rplayy·
@mattkratter @separ8 Is this only possible via Braiins? What about other hashers such as Riot or Cleanspark, can plebs rent their machines as well? But they probably have contracts with Foundry etc? Hope u find a minute to answer🙏🏻
English
1
0
3
210