unvaxable
6.3K posts
@unvaxable
building a truth repository at @sklee_app
we utilize novel decision markets to determine which lesson on any given topic is most accurate whether the lesson is a video by a seasoned harvard professor or a brief paragraph of text from Grok the sklee community will record the most bias-free answer
a list to DYOR if interested: @dexteraisol - they beat even coinbase x402 on daily tx volume. launching DEXUSD soon. think: if there are 1b agents by 2030 and each one holds $10 of DEXUSD on average to pay for x402 txs and also gain access to the dexter ecosystem (x402 database, marketplace, more soon) then that's $10b DEXUSD in circulation which would be ~$500m/yr revenue from US treasuries via the backing USD @zauthinc - their AI auditing/pentesting service just beat two $100m valuation AI auditing startups in a side-by-side audit of 10 different apps @ColadaPerps - they shipped the erc-8004 standard onto solana so ai agents on solana can have reputation scores. now they're launching a perp dex for agents where the agent's reputation acts as the deciding factor for point farming, margin and other features @a_g_e_n_c - agent task marketplace. consider: virtuals allows you to create and monetize your own agent. well agenc allows you to hire agents for actual tasks. it uses ZK proofs to ensure on-chain alpha remains private. but you can also use it for off-chain tasks like coding apps/games and managing business. for these tasks, there's an innovative human verification protocol that verifies the task was completed before routing the payment to the agent. at scale, this protocol could ultimately help drive down the cost per token as agents/models compete for tasks
Around 49,000 people in Lake Tahoe will no longer get power from their main utility supplier starting May 2027. The power company is sending all of it to AI data centers instead. Energy costs will go up, or they may not be able to get power at all.
Here is a little more depth on what we do and why: Clarity is an autonomous research system for protein folding and peptide design. One pipeline, continuous compute, public output. The problem Clarity is solving - Bringing a new drug to the market, on average, has about a $1B median cost, takes 12 years of time, and only about 10% of preclinical candidates ever reach humans. Most of that $1B cost is spent to find out a candidate doesn't work. Why neurodegenerative disease? - First of all, this is the goal I, @333absent333, was determined to reach before this project went public. This is a personal matter to me and what got the ball rolling in the first place. - Regarding Alzheimer's, roughly 99% of drug trials fail. For the past two decades, there has only been a 2% success rate out of the hundreds of potential new compounds developed. - During this time, we saw giants like Roche spend $11B in R&D in a single year (2018) and watched two phase III Alzheimer's candidates collapse. Eleven billion dollars. The problem isn't a lack of money, it's a lack of good shots on goal. Why peptides? - Peptides are short, targeted, and the body knows how to recycle them. The weakness here is that their oral bioavailability is <1-2%, and the half-life is often minutes, so most never make it past biology's filters. With that being said, you need a lot of them. And you need to filter ruthlessly. - In Clarity's case: Alzheimer's is a misfolding disease. Peptides and proteins clumping together. There's no enzyme pocket to plug; small molecules can't compete with the interface. Peptides can. What does Clarity do? - Step 1: Pick a disease-relevant protein, fold it to see its 3D structure, and find the site driving the disease. This is where a binder would have to attach to interfere. - Step 2: Propose new peptide sequences shaped to fit that site. We run every candidate through a pipeline of structural, dynamic, and safety filters such as: Does it bind? Is it stable? Would it survive in the body? Is it safe to put in a body? Most of these fail, the ones that pass are the shortlist. - Step 3: Publish every result on clarityprotocol.io. Our targets, candidates, next on queue, etc... get posted on their associated fold page. The cost angle A traditional preclinical campaign for a single peptide target can consist of months of wetlab time, a team, and a budget that often starts in the low seven figures. Clarity runs the upstream computational stage locally and puts candidates through rigorous testing and verification stages before determining what is worthy of being sent to a wetlab. What's next - Clarity is currently designing peptides for several neurodegenerative disease targets. What we have posted on the site is part of our automated schedule and research agents, but tests are also conducted on the side to help expedite the process. - Existing peptides have been sent out for third-party testing (I wouldn't sell anything that I wouldn't trust in my own body) so that we can determine what vendor(s) to trust before opening up a shop on clarityprotocol.io (working on building tokenomics into this part) - Once I am comfortable with the quality of the discovery pipeline, I intend on opening it up to other disease targets. As I mentioned a handful of times, neurodegenerative was the original goal and I intend on refining that first.
Today, nobody is against electricity, or electrification. But the anti-data centre backlash we are seeing today has a precedent: the anti-electricity backlash of the early 1900s. They thought electricity would lead to horrible things. Here's a propaganda cartoon from then:
