protocol design, web3 philosophy, internet capital markets, sovereign money systems | @SpreeFinance @CovenantLabsAI @DropChainCRM
USA
Joined November 2015
- Tweets992
- Following3.2K
- Followers2.6K
- Likes45.3K
We will see a precipitous uptick in 0-day exploits that are geopolitically motivated in the coming weeks/months
Not because AI is discovering them (though this is true and will accelerate)
But because highly funded nation state actors are sitting on a slew of 0-days, and this week has shown them that these attack vectors have an expiration date
When will @AnthropicAI / @OpenAI be sued for false advertising when they give performance metrics for a model at launch, then quantize it down but keep the same name/labels
Significant benefits to humanity through AI, like discovering new physics or cures for diseases, clearly isn’t worth it if some people lose their sense of identity
What are we even doing here smh
Carter retweeted
Access to Intelligence should be a permissionless MARKET, not a MONOPOLY
Meet Charis - the open market for encrypted AI
Live now on charis.im
This is a classic prisoner's dilemma. It assumes China will be truthful in it's disclosure of internal state secrets.
I'm exceptionally surprised that people are calling for a slowing to the AI race that has a very clear Nash equilibrium (hint: nobody slows down)
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Dario himself says that Anthropic would need to implement intrusive oversight:
"give outside evaluators desks, badges, laptops, internal tool access, employee-like permissions, access to training processes, and the ability to report publicly."
the equivalent to implement this into Chinese labs would be something like American/international inspectors receiving persistent access to China's most advanced AI laboratories, compute clusters, training runs, internal AI-research agents, experiment logs, model weights, military-linked programs, and perhaps classified research
I find it extraordinarily difficult to imagine the CCP agreeing to that level of access
Mass AI proliferation across social media will cause humans to talk and behave in ways that are indecipherable
Sarcasm, confusing in-group jokes, and natural language ciphers setup like English booby traps will be common to catch bots and evict them from the human commons
When these LLM botnet swarms execute their attacks and realize winnings, what will they store their value in?
I never would've guessed I'd support @sama and @OpenAI this quickly after being on the polar opposite of the spectrum not even a week ago
@DarioAmodei is doing an excellent job creating distrust by censoring frontier model use cases
Watching sci-fi movies published before LLM mass adoption
I find that the technologies like voice assistants “not recognizing a command” is a vestigial feature of a time before natural language interfaces
It’s increasingly hard to suspend disbelief during these movies knowing what is coming
oh my
I knew prediction markets would be huge for normie adoption... but this is blowing my mind.
Insurance markets on events means you can tailor your insurance policy around what you want to pay for, and not have to pay for a policy that is irrelevant to you (but is traditionally forced to be included within the policy by the issuer).
This creates a much more robust market for insurance bonds which will likely crowd out margin for large insurance companies... unless they start to embrace them right now and become a "UX layer" on top of these markets for the technically illiterate
Then again, a trained AI agent can do this for you... an Insurance Agent
When there is sufficient volume riding on a resolution for a prediction market, it increases the demand for (accurate) prediction resolvers.
There will be multi-billion dollar companies built around resolving predictions... this makes resolving events incredibly important to keep secure. I wonder if some separation of concerns / checks and balances should be standardized
Think Doordash but for events validation. "Earn $500 to travel to Bolivia and capture the amount of rainfall over the next 4 days". Must record the event at least 16 hours per day
huge — we are cooking up something great w/ @jaoudefi
We are proud to welcome Jaou Toure @jaoudefi as the new COO and Co-founder at Spree Finance.
Jaou brings deep expertise in stablecoins, digital assets, and financial infrastructure, and will help drive our mission to rebuild commerce and loyalty for the modern economy.
The industry is advancing quickly and we are excited to have Jaou help lead Spree into its next phase.
Our @spreefinance stable points system fixes this
hmu @GM
We just launched AIR.Shop tickets
40M+ concerts, games, and shows
buy a ticket, get up to 10% back in stable rewards, built with @Moca_Network + @SpreeFinance + @JoinBookit
I frequently find superior deals on air.shop vs existing web2 ticket providers, even before AIR SP back
If you’re in Asia, we are giving away BLACKPINK Tokyo Dome tickets, check it out: tickets.air.shop