Every time I think I’ve understood it, @gaib_ai shows another layer. Yield doesn’t just sit in contracts — it sparks, moves, reshapes the whole frame. It feels less like finance, more like standing inside a reactor.
@SuiNetwork@KRNL_xyz@LightLinkChain@dangoXchg@irys_xyz
major sports events usually become a stress test for platforms built around predictions and audience attention. it is during these periods that product retention and long term engagement become visible
Did you see yesterday’s move in $OPN from @opinionlabsxyz? And this is only the beginning
The 2026 FIFA World Cup is about to start, and the platform already has a huge number of prediction markets covering all kinds of events and outcomes
The excitement is likely to grow even more thanks to the new Regional Ambassador Program, featuring a leaderboard and a total prize pool of $50,000 running from June 3 to June 17 (all details are available in Discord)
We’re going to see even more prediction opportunities across different events
We’re just getting started...
it seems the strongest part here is not speed but simplifying the path of data movement.when extra coordination layers disappear the system gains advantages not only in latency but also flexibility
so, week started, i'm full of energy, even because i didn't pass selection for moderators at @ElasticsAI, but there's lot to say and i will continue
lost first match in new football tournament on saturday (future tournament champion 7-5)
i recently wrote about 400ms and many of you asked- how is this technically possible? figuring it out
in traditional oracles (chainlink, band) everything simple: nodes gather, vote, then result written to chain. it's slow, because you need to wait for consensus and transaction
@PythNetwork did it different way. publishers (jane street, wintermute) sign price right in their system. this signature flies to pythnet -separate blockchain, where prices aggregated. then app takes this price from pythnet through pull-model itself
it's like difference: wait for mail delivery (push) vs go to post office yourself (pull)
result: publisher signed price → it's on pythnet → app takes it → everything in 400ms
most interesting part: app pays for taking the price, not publishers paying for publishing. this changes whole oracle economy
that's why pyth can scale to 100+ blockchains without explosive growth of fees. bitmex uses this for rwa derivatives, kalshi- for prediction markets, polymarket - for traditional assets
not just fast. this is architecturally right
speed at four hundred milliseconds and the number of data providers only matter together with real use cases
the architecture truly handles load at this scale then this stops being a marketing story and starts becoming infrastructure
been digging into @PythNetwork network for a week and can't wrap my head around why nobody was talking about this earlier. at first it looked like just another oracle like chainlink, but then i saw who's here: jane street, wintermute, revolut just dumping data
138+ publishers updating 3,059+ price feeds in 400 milliseconds. not a second, 400ms. happening across 100+ blockchains simultaneously
as a developer - integration is 5 lines of code. as a skeptic - surprised it actually works. as someone tracking markets -this changes the whole DeFi infrastructure game
spinning up a 30-day deep dive:
why bitmex picked pyth pro for rwa derivatives
how kalshi (cftc-regulated prediction market) uses real-time data
the architecture that delivers this speed
tools on pyth api that'll find arbitrage
either this is just another marketing push - or i'm seeing something that matters. let's find out together
the acquisition pattern suggests streaming platforms are securing internal ai pipelines.control over generative production tools can reduce costs and accelerate content experimentation
good morning friends
if you noticed, more and more often the film industry uses AI tools to create shots or full scenes, including fully generated short films/cartoons/advertising, etc
for example, Ben Affleck had a company for film production using AI and the other day a deal was made to sell it to Netflix
i'm sure this is not the last acquisition of small companies by such a giant as Netflix
it's important to note that AI takes into account the database and the AI itself uses this data during editing/production/post-production
today it's always better to use AI trained on human data, as @getoro_xyz does, then the AI work process in cinema gets a huge advantage, in terms of what the viewer wants to see
rapid iteration between versions signals optimization cycles rather than architectural revolutions.the focus on reducing hallucinations suggests reliability
good morning friends, happy saturday!!!
interesting news
GPT-5.4 came out - just one day after GPT-5.3
what's main "under the hood"
fewer errors:
- probability of errors in answers decreased by 18%
- probability of false statements - by 33% compared to GPT-5.2
thinking mode became more transparent:
now the model shows a preliminary action plan so the user can guide it during the work process
the image was generated in the new version using a simple prompt, for the purity of the experiment and to see what it would do, generation happens in "thinking" mode:
"create an image of yourself as a new version and what advantages you have over the previous version, size 16:9"
the geopolitical view assumes rational cost benefit framing.which makes pressure tactics more plausible than immediate escalation unless incentives abruptly shift
After some time, we started talking a little. Do you think this is good or bad?
Let's move on to less important news, a little politics, and news about AI:
/// Trump, as usual, likes to stir things up in areas of US interest on Fridays, because that's when the stock market closes.
I agree with you on this. Will he decide to strike Iran? I don't think so. They did have negotiations, and it's clear that if things were really bad, it would have already happened. I think he just wants to use this pressure to achieve certain points that are in his interests.
P.S. I am sure that Iran is already prepared for a new round of hostilities and it will not be as easy as last time.
Well, there will be a counterattack.
/// Google has rolled out Gemini 3.1 Pro
The main highlight is logic and thinking:
77.1% on the ARC-AGI-2 benchmark for solving *unfamiliar* problems
For comparison:
- average person ≈ 60%
- previous Gemini 3 Pro (November) - 31.1%
That is, growth of 2.5 times in three months.
Where Gemini 3.1 Pro leads the way
- 94.3% GPQA Diamond - PhD-level scientific knowledge
- 80.6% SWE-Bench Verified - agency programming
- Elo 2887 in LiveCodeBench Pro - competitive coding
- 92.6% MMMLU - multilingual knowledge
Overall, the model outperforms Sonnet 4.6, Opus 4.6, and GPT-5.2 in almost every respect.
Google calls the model "basic intelligence," on which Gemini 3 Deep Think is already built (84.6% on ARC-AGI-2).
/// "Programmers will disappear by 2026" — an alarming prediction from the creator of Claude Code
Boris Chern, author of Claude Code from Anthropic, has stated that the profession of software engineer is on the verge of extinction. And not in decades, but as early as 2026.
On the Y Combinator Lightcone podcast, he said, "We will see the profession of 'software engineer' disappear."
According to him, the code will become fully automated, and people will:
- will write specifications
- communicate with users
- formulate tasks for AI
"Hawk Tuah girl"
the framing relies on contrast between spectacle and structured engagement.long term value however will depend on credibility and consistent delivery rather than rhetorical comparison.
Not For Anyone & Do Your Own Research
Has she returned? With no means to support herself, she must continue doing what she does well.
"Hawk Tuah girl" choose better @beeos_arenavs
the differentiation lies in separating model improvement from data extraction.using tee and cryptographic proofs reframes user contribution as controlled computation rather than silent harvesting.
ai @getoro_xyz and security of your data
oro does one thing that big tech fundamentally can't do: tries to separate ai training from your data leaks
the idea is this: your accounts and digital footprints remain "locked" in a protected environment the project claims it uses zktls and tee to pull data privately and process it so that what goes out is not the raw material, but the result: improved model, features, context
not your emails, not your dms, not your personal details - they remain locked
for the user this looks fair: you give access to training on your data, but don't give away the data itself as a commodity and if this approach scales, a new market appears: people become not "raw material" for ai, but suppliers of value on their own terms
@1BY1_UA
there is also a risk tradeoff being acknowledged here.lower cost structures reduce pressure to overpromise which helps align delivery with reality over longer cycles.
Indeed, real data will be more useful for small AI projects, as they are tailored to specific tasks.
There are mass-produced products, yes, undoubtedly, other capacities and mass production are needed there.
But @getoro_xyz has its own path that they have chosen, and I like this approach: less cost, more benefit, and greater productivity.
My focus isn’t just directed at this project
It’s laser focused. Yes, the same laser people turn on when they see real potential.
This isn’t a quick story. This is a long game.
With this mindset and this drive, I look at @opinionlabsxyz and make my choice in its favor.
> BNB Chain, CZ and prediction markets. Do you feel the potential?
Get ready for big events and new heights.
gOpinion
so, the beginning of the month and week opens with @vooi_io for us
- personally i don't understand the reason for extending registration, it could be a loyal step towards users
- or maybe there were supposed to be 19 epochs, but the team decided to make all 20 epochs
- or it's both options that i mentioned above, because the team saw the market situation
basically it's unclear to me personally
i also hope to see news not only about the new epoch and weekly breaking of ATH in trading volume
i would like to see details about TGE and snapshot (obviously for me it will be done next week)
When is the listing $VOOI?
You simply can't imagine the experience I got from @almanak
To say I was blown away is an understatement
Considering that I understand how AI works, it's simply superb
In short... it's 4 figsssssss and it's a trial run
the key point is that the author stresses the gap between interest in basic finance and the lack of attention to tools that actually produce returns.almanak is mentioned not for hype but as an example where strategy and automation already function with practical usefulness.
so what’s happening again
some of you forgot or never knew that real money is made in defi
why not explore that
why are you ready to learn how loans work
but not how your own money moves
why are you still ignoring ai
the truth is
things move way faster than before
way faster
why not take a closer look at @almanak and their tech + ai
i’m telling you this thing will take off harder than spacex
everything in its time
but while you ignore it
someone else is deep in almanak and making money
like me and a whole bunch of others
the author highlights the gap between noisy marketing and concrete mechanics and the key detail is the daily reimbursement which keeps entry low and interest steady. it becomes a clean way to gather user signals and engagement. it makes sense to watch
i like some of the approaches people use to promote their products
some shout that community is everything and try to be like the $mon token but can’t even raise enough for the sale
i’ve said it many times already you just can’t do that
here’s a perfect example
there are other projects that don’t scream about community
they just build cool things
not always popular
but interesting
like @useTria for example
they launched a fun campaign
they’ll pay for your purchase receipt
main conditions
have their card
make a post in x
attach your receipt preferably
tag them and add #triatreasure
once a day they’ll pick someone and cover it
looks fun at the very least
go pay for your lambo and upload the receipt
wouldn’t be surprised if someone actually gets it paid back
maybe that someone is you
again this whole thing
investors are sounding the alarm - the risk of an AI bubble…
the surveyed fund managers control over $475,000,000,000 in assets (november Bank of America survey)
▫️ key risks: "AI bubble" (45%), rising government bond yields (17%), return of inflation (16%).
▫️ 53% of investors consider the AI sector a bubble, even though artificial intelligence is also called the main driver of company growth.
▫️ a record 63% of respondents believe the global stock market is overvalued, yet fund managers still increase the share of stocks in their portfolios.
and so on
i still think this is one of the ways to "heat up" the sector
the time has come where AI will provide analytics better than humans
that’s why @GlintAnalytics uses it in this sector
this isn’t stupidity, it’s not following a trend
these are real steps that must be used, and they are doing it
no matter how much you want to follow their logic, forget it, this is news created to control your opinion
do your own analytics
your tone is steady ok and the contrast with the messy payment flow makes the point stronger
the key is whether tria has a stable routing layer otherwise convenience evaporates fast
try running all small payments through it for a month to see where delays appear
played football today, third time this week, it’s great, lately not many people show up
usually we play once a week
and it often happens that for renting the field you have to pay either in cash, or by card, or different banks
honestly it annoys me
and i think many people feel the same, it’s sometimes just terrible
since i discovered @useTria, they solve this problem, one wallet, tons of networks, pay how you want and on any chain
i really want banks to connect to their system
so this can be solved instantly and with cheap fees
agree, you want the same thing too
and the fact you started checking glint stats right away shows the right mindset here it matters not only how the interface looks but what logic they actually offer the user if the product aligns with what you already use in infofi the synergy will show up naturally without
and so, i’m getting my guaranteed allocation for making the top 100
this is important for me personally, you could call it my first real result in infofi
i understand there’s still a lot to do, and it’s important to do it consistently, improve along the way, and admit mistakes - that’s a must
i applied for the @GlintAnalytics whitelist
for now i can look at some of the available stats
i need to understand how the project works, and i always use a product if i talk about it
you should never speak about something you don’t use or aren’t even familiar with
that’s a road to nowhere
see you tomorrow my favorite activists
a friend of mine spent weeks trying to rent GPU power for his startup
banks wanted collateral, investors wanted traction, providers wanted a track record
he had none of that - just a model that actually worked
three months later he raised liquidity through tokenized compute resources
no banks, no paperwork, no intermediaries - just real assets verified and launched via @gaib_ai
now instead of endless talks he manages his own compute pool and earns from AI tasks
the irony is that access to GPU didn’t come from corporations, but from DeFi - a system that finally started working for real people