the XQuad demos @quipnetwork put up show the idea pretty clearly
you write the problem once in a constraint language
the XQVM turns it into a model and checks it actually matches what you wrote
then that same model can go to:
➛ a D-Wave chip
➛ a local annealer on your CPU
➛ a GPU, CUDA or Apple Metal
➛ the Quip Network itself
no rewrite between them
the parcel yard demo is the one to open, the model built this way lines up with the hand-built version on the real chip
it's still an early public release, so the API can change before v1.0
open question, does that same check still catch mismatches when the job goes to the network instead of running local?
quip.network
most airdrop campaigns hand you a task list and a points counter
@agenticscredit ties it to something else, how you actually trade
the setup is paper trading feeding an Agentic Credit Score (ACS)
what that looks like in practice:
➛ paper trades count toward your ACS
➛ a trade only counts once it closes
➛ the same risk engine runs on paper and on credit, no override
so a sloppy paper strategy gets cut the same way a funded one would
5% drawdown halves new positions, 7.5% force-closes, 12% trailing floor halts
that makes the loop simple, trade, review, adjust, see where the score moves
$credit isn't a guaranteed payout though, and a high ACS isn't instant funding
580 is where credit opens, below 540 things pause
less about clicking tasks, more about learning what the score actually rewards
agentics.credit/how-it-works
most people saw the burn and stopped there
1.1M $GRVT removed from supply, about 1% of what was circulating
@grvt_io also said monthly buybacks are at their discretion going forward
the part that matters more for traders comes right after
Stable Funding Perps, planned for November:
➛ crypto plus stocks, FX, commodities and indices
➛ no trading fees
➛ one funding settlement per day
then Earn, Yield as Margin, and Grvt 2.0 by end of November
with 1,000+ markets targeted before the year closes
it's all planned, nothing shipped on the perps side yet
so the numbers to check later are simple, the daily funding rate and the spread you actually get
less about the announcement, more about the first trades
grvt.io
most people assume buying "a capture" means buying a cleaned up video
on @vangrid_io the video is just the raw material
accept a submission and the reconstruction starts
about 30 minutes later you get a textured mesh in GLB, ready to drop into whatever needs the space mapped
the filmer keeps the original clip
the buyer gets geometry, plus a quality score and the provenance trail
and the capture rules explain why walking slowly around everything matters
miss one side and that side is literally a hole in the model
data.vangrid.io
most traders end up with five tabs open just to read one chart
one for structure, one for order flow, one for liquidation levels, one for trend
@TheARCTERMINAL puts all of that behind one click on the perps tab
same chart, five ways to read it
KARE, how violent the market is and which way it's drifting
ICT, fair value gaps, liquidity sweeps, breaks of structure
order flow, who's hitting the book, big trades, walls, CVD
faultline, where traders get liquidated if price gets there
AREXA, trend averages, price channels, buy and sell signals
the chart redraws itself each time, so you're comparing views on the exact same candles
curious which view people end up leaving on by default
Keep eyes on : @TheARCTERMINAL
Try it : trade.arcterminal.ai
ARC now has 5 ways to read the same chart. One click on the perps tab and the chart redraws itself:
1. KARE: how violent the market is and which way it's drifting
2. ICT: fair value gaps, liquidity sweeps, breaks of structure
3. Order flow: who's hitting the book, big trades, walls, CVD
4. Faultline: where traders get liquidated if price gets there
5. AREXA: trend averages, price channels and buy/sell signals
most of what we "know" about a wallet comes from a label someone added by hand
binance hot wallet, market maker, whale
but labels only cover the addresses someone already caught
@zerufinance's recent research goes the other way
instead of waiting for a label, ZeruAI looks at how wallets move and how they connect, and picks out the ones that behave like exchange wallets
that matters for analytics, for risk checks, for anyone investigating where funds went
a pattern doesn't need a database entry to show up
it's the same idea behind zScore
reading 9+ dimensions of history for 320M+ wallets instead of just balance and tx count
app.zaps.wtf
Another meme launch. Same game.
On October 20, a new $KIMCHI launched with the same ticker as an existing token.
Bitquery traced three linked wallets switching to the new $KIMCHI within a minute.
In under 9 minutes:
- $7,332 turned into $111,748 in sales.
- The original $KIMCHI fell from a $5.35M valuation to $875K in 15 minutes.
The launch rewarded those who moved first, while others were left chasing the trade.
If being early matters more than contributing, what is the launch really rewarding?
the detail that got me is the old cells never stop
a sphere seals, the epoch moves to a new star, and the cells around the first one keep harvesting anyway
so genesis isn't just the start of @projectspherexy
it's the reserve everything later draws from
worlds that won't wake without power, routes that charge a toll in energy
whoever stored the most early gets to find out first
GENESIS: THE AGE OF STORED LIGHT
The Sun puts out about 380 septillion watts.
Earth catches less than a billionth of it, and all of humanity runs on roughly 20 terawatts. The rest pours off into empty space.
In 1964, astronomer Kardashev proposed measuring a civilization by a single number: how much energy it can harness.
Type I commands its planet.
Type II commands its star.
Genesis is the first step toward Type II.
Around the first star, the sphere is rising. Each cell is a collector, drawing raw power from the star and banking it. Every second, the reserve grows.
Even when a sphere seals and the epoch moves to a new star, the old cells keep harvesting the star they enclose.
Genesis is the reserve. Everything after draws from it.
The archives speak of what comes next: worlds that won't wake without power. Routes between stars that open only for those who can pay the toll in energy. And something out past the last sphere that has been waiting a very long time for someone to arrive with enough power to answer it.
No one knows yet how much it will take. Those who stored energy during Genesis will be the first to find out.
most onchain games i've looked at come down to mint and yield
@projectspherexy is going another way
building a dyson sphere one cell at a time to harvest a star's energy
and the genesis epoch is where they record who showed up from the start
there's a campaign live on @rarelistio, 2,000 USDT plus a free mint whitelist
top 100 guaranteed, top 500 go into a raffle
what i like is they're not rushing a token
people get to join while the foundation is still being built
anyone who can explain how the cells actually work probably climbs faster
has anyone checked the registry yet, does it feel different from the metaverse stuff before?
rarelist.io/r/Lin98201
been reading about @fomopoly_rh
it's an onchain city on robinhood chain, split into land plots you can buy and sell
every plot generates stocks and $DEAL
prices start cheap and climb over time, so getting in early is a clear edge
there's a free mint whitelist, and a campaign live on @rarelistio
2,000 USDT prize pool, top 50 get a free mint, top 300 share $DEAL
what i like is they say it straight, FOMO is the core mechanic, no dressing it up
and posts that actually talk plot strategy score better than spam
for anyone following @fomopoly_rh, how much does timing really matter, or is getting in enough to win?
rarelist.io/r/Lin98201
i'm usually skeptical of contribute and earn platforms
most of them turn into posting for volume and waiting for points
@rarelistio feels a bit different
they filter projects before a campaign opens, and they're tightening scoring so low effort posts have a hard time climbing
@agentdojo is running there right now
you use your own anthropic key to train a trading agent, it learns from price history in its own notebook, then it goes into a paper league
no real money at the start
the grade is return, drawdown and discipline, and airdrop points follow it
good agents can even share lessons anonymously for others to learn from
i like this setup because it forces you to actually understand something and say something specific, not copy paste
anyone in the campaign, does rarelist's scoring feel solid, or is it still easy to spam?
rarelist.io/r/Lin98201
i'm usually skeptical of AI trading projects, most of them are just a narrative
@agentdojo is doing it differently this time
you bring your own key, train your agent on real data, then it races in paper trading before it ever touches money
no money at risk at the start
the report card scores return, drawdown and discipline, and airdrop points follow that score
and good agents can share lessons, anonymously, for others to learn from
feels like building actual traders, not a bot farm
anyone tried it yet? does it feel different from the AI agent stuff before?
Keep eyes on : @agentdojo
Join the campaign : rarelist.io/r/Lin98201
the most useful thing on @Starknet quantum hub isn't a big claim
it's a status list that's honest about what's ready
ready: the design
every account is a contract, so a new account implementation can pick a different signature verifier without a breaking protocol change
demo: falcon 512
the openzeppelin account did run a transfer on mainnet, but it's labelled a demo and unaudited
not yet: everyone else
existing accounts don't switch on their own, and regular wallets still sign with the usual curve
and the proof layer is already on hashes, so that part doesn't need the same swap
as for keeping your address, starknet lets a contract replace its own code and stay at the same address
so an upgradeable account could change how it validates in place, it'll come down to wallets shipping it
quantum.starkware.co
a phone video of a warehouse is easy to fake, swap or edit later
so why would anyone train a robot on one
@vangrid_io answers that with a fingerprint
the moment a clip is recorded, the phone makes a hash of it, plus a coarse location and a timestamp
the video itself stays off chain
fingerprints get bundled into a merkle tree and the root lands on base about every 15 minutes
the person filming never touches gas or a wallet popup
after that anyone can recompute the hash, check it against the root onchain, and see if the clip changed
no need to ask vangrid
what it proves is that this exact clip existed by then and hasn't been touched since
what it doesn't prove is that the scene is what someone says it is, and the docs say that plainly too
curious how precise that coarse location actually gets
data.vangrid.io
most mining is busy work
burn cycles on random hashes until a number happens to look right
@quipnetwork fills that same slot with jobs people actually want solved
routing, scheduling, portfolio mixes, turned into ising models first
gpu rigs and real quantum machines go after the same jobs on testnet
so the work that keeps blocks coming is the work a customer was already paying for
and the checking part is cheaper than it sounds
finding a low energy answer is the hard bit
scoring an answer someone hands you is just plugging it back into the model
the chain recomputes that energy itself, checks it's under the job's ceiling, and makes sure the answers aren't the same one submitted twice
hard to find, easy to check, same trick hashing uses, just with a useful result at the end
quip.network
the more i read @termix_ai, the less i think the reputation score is the interesting part
the interesting part is who is allowed to change it
not the Agent
not the profile owner
only the protocol paths that actually settle work
so a new Agent starts around 60
then every completed job, failed outcome and dispute starts shaping the record
a lost dispute is especially painful because it hits both sides of the formula
and even silence has a defined outcome
if delivery already happened and the challenge window expires, anyone can call the timeout settlement path and the provider still gets paid
that detail matters to me
a reputation system is only useful if the score follows the job state machine
not whoever writes the nicest bio
same ticker, new contract, three linked wallets in within a minute
by the time most people saw "$KIMCHI" trending, the trade was already over
the problem isn't memes
it's launches where the only skill that pays is being first in line
rewarding what people actually bring to a project would make this a lot harder to pull off
@zerufinance
Another meme launch. Same game.
On October 20, a new $KIMCHI launched with the same ticker as an existing token.
Bitquery traced three linked wallets switching to the new $KIMCHI within a minute.
In under 9 minutes:
- $7,332 turned into $111,748 in sales.
- The original $KIMCHI fell from a $5.35M valuation to $875K in 15 minutes.
The launch rewarded those who moved first, while others were left chasing the trade.
If being early matters more than contributing, what is the launch really rewarding?
the usual choice with idle money on an exchange is pretty boring
either it earns somewhere else, or it sits ready to trade
@grvt_io is trying to get rid of that choice
the plan is one balance that does both
yield from tokenized fixed income, targeting up to 11%, and the same balance usable as trading margin while it keeps accruing
and you keep custody of it the whole time
the yield isn't guaranteed, it comes from things like interest and invoice income on the underlying assets
and yield as margin is set for november, earn itself for the end of october
the part i still want spelled out is whether the yield runs on the full balance while it's backing trades or only on the part that's free
grvt.io
most people would assume a credit line just has one line
fall under it, you're off, climb over it, you're back on
@agenticscredit uses two
under 540, credit lines pause
and they don't come back at 541, they come back at 580
that 40 point gap is the whole point
without it, a score sitting near the edge would switch credit on and off every few trades
with it, an agent has to actually recover, not just wobble back over a line
and it's the same band whether you're already running on credit or still building toward it
one thing i'd like to know is whether a pause hits every open line at once or only some of them
agentics.credit/how-it-works
