TradingGain-X

Sometimes I look at systems like this and honestly don’t know whether I’m looking at the future… or just a smarter version of old problems.
That thought kept coming back while reading through @OpenLedger and trying to understand what they are actually building around AI contribution, validation, and ownership. Because at first glance, it looks like another decentralized AI narrative. Data. Models. Incentives. Everyone says those words now. But the strange part is… OpenLedger feels less obsessed with “maximum freedom” and more obsessed with controlled usefulness.
And honestly… maybe that’s the uncomfortable direction AI eventually has to move toward.
People love the romantic idea of decentralization where everyone contributes anything at any time without restrictions. Sounds beautiful in theory. But AI systems don’t really survive on freedom alone. They survive on signal. Cleanness. Relevance. Trust. Without that, everything slowly turns into noise pretending to be contribution.
That’s why the Datanet structure caught my attention more than I expected. Daily upload caps. Validation rules. Text separated from image data. Audio isolated into its own lane. At first it sounds restrictive, but maybe restriction is exactly what stops these systems from collapsing into meaningless quantity farming.
Because we’ve already seen what happens on the internet when incentives only reward volume. Spam becomes strategy. Noise becomes economy.
Here though… acceptance rate matters more psychologically than raw output. That changes contributor behavior in a subtle way. You stop thinking “How much can I upload?” and start thinking “Will this actually be useful?” And weirdly, rejected contributions don’t feel like punishment inside that system. More like filtration. Experimentation still exists. The network just refuses to pretend every contribution has equal value.
Maybe I’m wrong here, but I think that distinction matters more than people realize.
The other thing that stayed in my head was ModelFactory. Not because GUI-based fine-tuning is revolutionary by itself, but because it quietly changes who gets to participate in AI development. For years, training models felt culturally locked behind terminal windows and highly technical workflows. Almost like complexity itself became a gatekeeping mechanism.
OpenLedger seems to be softening that edge without completely removing structure. Which is important. Total accessibility without safeguards usually creates chaos. But too much control kills experimentation.
And this is where LoRA and QLoRA actually make practical sense. Finally someone is talking about realistic AI iteration instead of pretending everybody needs infinite compute for full fine-tuning. Real people need workflows that allow train → test → interact → refine without turning their laptops into melting machines.
The broad ecosystem support is interesting too. LLaMA. Mistral. Qwen. DeepSeek. BLOOM. GPT-2. On paper it looks like feature expansion. But honestly, I think it’s more philosophical than technical. Diversity of models creates diversity of thinking styles, response structures, biases, strengths. AI ecosystems become healthier when experimentation isn’t trapped inside one dominant architecture.
There’s also this strange “disciplined kitchen” feeling throughout the whole system.
Contributors cannot just throw random ingredients everywhere and call it valuable. The kitchen still has standards. Validation. Measurements. But at the same time, everyone can taste the outcome, evaluate quality, and refine the recipe collectively. That balance between openness and discipline feels incredibly hard to achieve.
Even the documentation feels different somehow. Not like static docs people skim once and forget. More like a queryable intelligence layer where the system itself slowly teaches you how it wants to function.
And maybe that’s the deeper question underneath all this.
Will data truly become an asset class in the future… or are we simply renaming old trust and validation problems with newer infrastructure and better interfaces?
I genuinely don’t know yet.
But I think systems like @OpenLedger are forcing people to confront that question more honestly than most projects do.
#OpenLedger $OPEN

#OpenLedger @OpenLedger
A while ago I started noticing something strange whenever people talked about AI breakthroughs. The conversation almost always centered around the model itself. The company. The valuation. The benchmark scores. Maybe the founder if the story needed a face.
But almost nobody talked about the thousands of smaller contributors hiding underneath the surface.
The researchers whose papers quietly shaped the architecture. The communities producing open datasets. The people labeling data for hours without visibility. The testers finding failures before public release. The developers building tools around systems they don’t control. Even ordinary users, honestly, who interact with these models every day and unknowingly help refine them over time.
AI feels collective when you look closely enough.
And yet recognition inside AI still feels extremely centralized.
That part stayed with me.
Because the more powerful these systems become, the stranger this imbalance starts to look. Intelligence is being assembled from enormous layers of human contribution, but once those contributions are absorbed into a large model, the memory of who helped build it begins to disappear.
Maybe that is the real hidden problem in AI. Not capability. Not scale.
memory.
Most AI systems today are very good at generating outputs, but surprisingly bad at preserving contribution history in a transparent way. You can see the product. You usually cannot see the chain of people who helped shape it. The labor becomes abstracted away behind the interface.
And once contribution becomes invisible, accountability becomes blurry too.
The strange thing is that contributors are starting to realize how much value they actually create. You can feel that cultural tension growing slowly underneath the AI industry. People no longer want to simply participate in systems they cannot see into. They want provenance. Attribution. Context. Some kind of permanent record that acknowledges their role.
Not necessarily ownership in the traditional sense either. That distinction matters.
Ownership is rigid. Participation is fluid.
A researcher might contribute an idea. A community might improve a dataset. A developer may fine-tune infrastructure. Another person may identify failures the original creators missed. AI development increasingly looks less like a single invention and more like an evolving network of collaboration.
But collaboration without memory creates imbalance.
This is partly why blockchain keeps reappearing in AI conversations, even when people are skeptical of the hype around “AI + crypto.” The interesting part was never simply attaching tokens to machine learning. It was the possibility of creating transparent records around contribution itself.
That is honestly why @OpenLedger caught my attention.
Not because it promises some magical future where blockchain fixes AI overnight. Most people are tired of hearing that. What feels more interesting is its focus on attribution and traceable collaboration inside AI systems. The idea that AI networks may eventually need a memory layer capable of recording who contributed what, when, and how.
That feels more important than people realize.
Because trust in AI probably won’t come only from smarter models. It may come from systems that can openly show their history instead of hiding it behind corporate walls. Systems where contribution can be verified rather than assumed.
And maybe that changes behavior too. Maybe transparency naturally creates healthier ecosystems because people know their work cannot quietly disappear into centralized infrastructure without acknowledgment.
I keep coming back to the same thought lately: intelligence alone is not enough.
If future AI systems cannot remember the people who helped create them, then the imbalance we see today will only deepen as the technology grows more powerful. But if systems begin to preserve contribution, provenance, and collaboration in a transparent way, then AI could evolve into something far more accountable than what exists now.
Not perfect. Probably never perfect.
But at least honest enough to remember where intelligence actually came from.
$OPEN

A while ago I started noticing something uncomfortable about the way people talk about artificial intelligence. Whenever a model becomes successful, the conversation usually narrows around the company behind it, the founders, the valuation, the infrastructure. But the deeper I looked, the harder it became to ignore how many invisible hands are buried underneath every “intelligent” system we use.
Someone cleaned the data.
Someone labeled edge cases for hours.
Someone tested outputs nobody else wanted to examine.
Researchers shared ideas openly.
Communities refined models through constant interaction.
Developers built tools on top of imperfect foundations.
And then somehow all of that work dissolves into a single branded product.
That part stayed with me.
AI systems are often described as if they emerged from pure computation, but they are really layered accumulations of human contribution. Thousands of small inputs become compressed into something that later appears seamless and centralized. The strange thing is that once intelligence absorbs those contributions, the memory of where they came from begins to disappear.
I think this is one of the least discussed problems in AI right now.
Not scaling.
Not inference costs.
Not even regulation.
Memory.
Or more specifically: attribution.
Because intelligence without memory creates imbalance....
The current structure of AI rewards ownership far more than participation. The people closest to the final model capture most of the visibility and economic value, while contributors further down the chain slowly become anonymous. Data providers become invisible. Testers disappear. Smaller researchers get absorbed into larger ecosystems without much acknowledgment beyond a citation buried somewhere nobody reads.
Maybe that was always inevitable. Large systems simplify history. They compress complexity into cleaner narratives. But AI feels different because contribution itself is becoming the raw material of the industry.
The more capable these systems become, the more valuable human input becomes too.
And contributors are starting to realize that.
You can already feel a quiet cultural tension forming between centralized AI companies and the people feeding these systems. Not always openly. Sometimes it shows up as frustration over scraped content. Sometimes through debates around ownership, compensation, or dataset provenance. Sometimes it appears in smaller ways, like creators questioning why their work strengthens systems that never remember them afterward.
I don’t think people only want payment.
I think they want recognition that their contribution mattered.
That’s a different thing entirely.
This is where blockchain starts becoming interesting to me again, although probably not in the way most crypto narratives frame it. The important part is not speculation or branding everything as decentralized. It is the possibility of creating persistent memory around collaboration.
A transparent record layer.
Something that can track contribution instead of erasing it.
When people hear “AI + blockchain,” the conversation usually drifts toward tokens very quickly. But the more meaningful question may be simpler: how do future AI systems remember who helped create them?
Because if AI becomes deeply collaborative, provenance stops being optional.
You need systems capable of tracing where information came from, who improved what, which datasets influenced outcomes, and how intelligence evolved across time. Not just for fairness, but for accountability. Trust becomes difficult when the creation process is hidden behind centralized walls.
And honestly, transparency changes behavior naturally.
People act differently when contribution histories are visible.
Communities become more careful.
Collaboration becomes easier to verify.
Credit becomes harder to monopolize.
That does not magically solve exploitation or imbalance. But it changes the structure of memory itself.
That’s one reason @OpenLedger caught my attention recently.
Not because it presents itself as another “AI crypto” project. We already have enough of those. What feels more interesting is the underlying direction it seems to explore: the idea that collaborative AI systems should be able to remember contribution instead of absorbing it into silence.
There’s something subtle there that I think many people are overlooking.
OpenLedger appears less focused on replacing AI companies and more focused on building infrastructure around attribution, provenance, and transparent participation. almost like an attempt to give AI systems a historical layer they currently lack.
And maybe that layer becomes essential later.
Because the future of AI probably cannot depend entirely on blind trust in centralized actors. The systems are becoming too influential, too integrated into decision-making, too economically important. If contributors cannot trace how value moves through these ecosystems, resentment eventually builds underneath the surface.
History usually matters more than people expect.
Especially in systems built from collective intelligence.
The strange irony is that modern AI has extraordinary computational memory while still failing at social memory. Models can retain patterns across unimaginable scales, yet the humans who shaped those patterns often disappear from the story almost immediately.
That imbalance feels unstable to me.
Maybe the next stage of AI is not only about making systems more intelligent..
Maybe it is about making them more accountable to the networks of people who helped create that intelligence in the first place.
Because in the end, intelligence alone is not enough.
A system also needs the ability to remember.
#OpenLedger $OPEN

The first time I seriously thought about AI from a blockchain perspective, I wasn’t thinking about tokens or market hype. What stayed in my head was something much smaller and more human.
I noticed how AI systems often sound like they were built by a single intelligence, even though they are actually shaped by thousands of people working quietly in the background.
Someone labels data for hours.
Someone tests edge cases nobody notices.
Someone contributes open-source tools that later become foundational.
Someone provides niche datasets that improve models used by millions.
And then, somehow, those people disappear from the story.
That part stayed with me.
The strange thing about modern AI is that it has intelligence, but very little memory. It absorbs contribution without preserving visibility. Once knowledge enters the system, individual traces become difficult to see. The model becomes smarter and more valuable, while the people who helped shape it fade into the background.
Maybe that is the hidden tension underneath the current AI boom.
Because contributors are starting to realize they are not just users anymore. They are part of the creation process itself.
For years, AI companies collected data, feedback, testing, and interaction almost invisibly. Most people accepted it because AI still felt experimental and exciting. But now these systems are becoming economically powerful, and the imbalance feels harder to ignore.
Who owns intelligence that was built collectively?
Not philosophically. Practically.
That question keeps appearing everywhere — in discussions around datasets, creator rights, synthetic media, and open-source AI. People are beginning to care less about raw capability alone and more about where that capability came from.
Because intelligence without memory creates imbalance.
The system remembers patterns but not people.
It remembers outputs but not origins.
It remembers usefulness but not contribution.
Centralized AI systems make this worse because contribution histories are usually hidden behind private infrastructure. Most users cannot trace how models evolved or whose work shaped them. The architecture itself hides provenance.
And over time, that opacity erodes trust.
Not immediately. Slowly.
At first, people contribute because the technology feels exciting. Later they begin asking where the value flows. Eventually they ask who gets remembered once the system succeeds.
That is where blockchain becomes interesting to me again — not as speculation infrastructure, but as a transparent memory layer.
Not ownership in the simplistic “everything becomes a token” sense. More like accountability through traceable participation.
There is a meaningful difference between owning a system and contributing to its intelligence. Traditional AI platforms often blur those lines. If you do not own the infrastructure, your contribution becomes temporary while the platform keeps the long-term value and memory.
But transparent systems change the emotional dynamic entirely.
When contribution becomes visible, people stop feeling like raw material and start feeling like participants in the architecture itself.
That distinction matters more than we think.
It is also why projects exploring transparent AI collaboration are beginning to attract attention. Not because people suddenly love buzzwords again, but because the underlying problem feels real.
@OpenLedger is one of the more interesting examples I’ve seen recently because its focus feels different from the usual “AI + blockchain” narrative. What stands out is not simply the combination of technologies, but the idea of remembering contribution inside collaborative AI systems.
How do you track who contributed what?
How do you preserve provenance as models evolve?
How do contributors remain visible after systems become commercially valuable?
Those questions feel increasingly important.
Future AI systems probably will not be built by isolated labs alone. They will emerge from distributed networks of developers, researchers, communities, data contributors, evaluators, and users. Intelligence is becoming socially constructed at scale.
But social construction without attribution creates extraction.
And people can feel that instinctively.
There is also a cultural tension forming between centralized AI companies and contributors themselves. Companies need constant streams of data, feedback, and interaction. Contributors are becoming more aware that those inputs create enormous value.
Yet the systems connecting the two sides remain mostly opaque.
That opacity may become one of the biggest trust problems in AI.
Because transparency changes how communities relate to technology. Even small forms of attribution create accountability. They acknowledge that intelligence is relational — that powerful systems do not emerge from nowhere but from layers of human cooperation...
Sometimes I think people talk about AI as if it arrived independently from society. But underneath the abstraction, AI is deeply human. Every model contains invisible fingerprints from millions of interactions, corrections, judgments, and contributions.
The intelligence feels synthetic.
The contribution history is profoundly human.
And maybe the future health of AI depends on whether systems can preserve that history instead of erasing it.
I keep returning to that idea because memory shapes fairness more than we admit.
If systems cannot remember contribution, power concentrates around whoever controls infrastructure.
If systems can remember contribution, participation becomes harder to erase.
That will not solve every problem. Human systems are too messy for simple solutions. But transparent contribution layers at least create the possibility of accountability, and right now that possibility feels important.
Especially in a world where AI influence is growing faster than public understanding.
The deeper I look into this space, the less interested I become in loud promises about artificial superintelligence or trillion-dollar markets. What interests me more is whether future systems can remain socially legible. Whether people can still see how intelligence was built. Whether contributors remain visible after value compounds.
Because intelligence alone is not enough.
A system also needs memory.
It needs provenance.
It needs a way to acknowledge the people whose invisible work made the intelligence possible in the first place.
That may become one of the defining questions of the next AI era.
Not simply whether AI becomes more powerful.
But whether the systems we build can still remember the humans inside them.
@OpenLedger #OpenLedger $OPEN

That’s been my main takeaway after spending the last year bouncing between chains, testing DEX routes, farming random incentives, and trying every “AI-powered” dashboard people keep posting about. Most of them stop at information. They summarize sentiment, scrape wallets, explain narratives, maybe point out trends earlier than Twitter does. Useful sometimes, sure. But when markets move fast, information alone rarely matters. Execution does.
That’s why @OpenLedger and the Octoclaw launch caught my attention more than I expected.
Not because I suddenly think AI agents will magically replace traders. I don’t. Honestly, most agent narratives still feel early and a little overhyped. But Octoclaw seems focused on something more practical: reducing the friction between decision-making and execution across chains.
And if you trade on-chain regularly, you already know friction is half the battle.
A lot of profitable setups die during the process itself. You bridge assets, wait longer than expected, gas spikes hit at the wrong moment, approvals fail, liquidity shifts before execution finishes, or some route suddenly becomes inefficient. Sometimes the opportunity cost hurts more than the actual loss. I’ve had trades where the thesis was correct but execution across chains completely ruined the entry.
That’s the part people outside DeFi underestimate.
The interesting thing about Octoclaw is that it’s leaning toward autonomous action instead of passive assistance. There’s a big difference between AI telling you something and AI actually carrying out a sequence of transactions efficiently across multiple environments.
I think timing matters here too.
A few years ago, this kind of infrastructure probably would’ve broken under its own complexity. Bridges were unreliable, account abstraction barely existed in practice, and modular systems weren’t mature enough. Now the tooling is noticeably better. Cross-chain infrastructure still isn’t perfect, but it’s far more usable than during the last cycle.
That doesn’t mean there aren’t risks.
If an AI agent eventually gets permission to move capital, route trades, rebalance positions, or interact with contracts automatically, security becomes the entire conversation. Permissions, wallet architecture, exploit surfaces, contract vulnerabilities — those issues don’t disappear because the interface feels smarter. In fact, they probably become more important.
So I’m still skeptical. I think everyone should be.
But I also think the direction makes sense.
If autonomous systems improve enough, speed advantages could eventually become commoditized. Right now, experienced on-chain traders still gain edge from reacting faster manually, navigating liquidity quicker, or understanding cross-chain flows better than average users. But if agents can execute those mechanics instantly, the edge may shift somewhere else entirely.
Maybe future advantage comes from strategy design instead of execution speed.
Maybe traders spend less time clicking buttons and more time designing frameworks, risk models, and conditional logic for agents to operate within.
That’s where @OpenLedger starts looking less like “another AI crypto project” and more like infrastructure.
And infrastructure usually becomes most valuable when users stop noticing it.
Nobody logs in excited about TCP/IP or cloud routing layers. People care because everything works smoothly on top of them. If Octoclaw eventually reaches that kind of invisible reliability for cross-chain execution, that could matter more than any short-term narrative around AI agents.
Still early. Still plenty of unanswered questions.
But compared to most AI stories in crypto right now, this at least feels connected to an actual trading problem.
$OPEN #OpenLedger

been thinking a lot about why some AI chains get attention fast… but developers still don’t actually build on them.
that gap matters more than people think.
A lot of projects talk about AI agents, decentralized models, data ownership, Proof of Attribution, or autonomous economies. cool ideas. seriously. But developers are usually more practical than narratives. The first thing most builders ask is way simpler:
“Can I deploy without relearning everything?”
That’s why OpenLedger’s EVM-friendly approach stands out to me.
@Not every blockchain problem needs a brand new programming model. Sometimes the smarter move is reducing friction instead of reinventing every layer. And honestly… that seems to be what @OpenLedger is trying to do with its architecture around Ethereum compatibility, OP Stack, EigenDA, and Ethereum settlement.
The AI narrative in crypto has become crowded very quickly. Every week there’s another protocol promising decentralized intelligence, model ownership, data monetization, or AI-native infrastructure. But a surprising number of these projects accidentally create too much complexity for builders.
They introduce unfamiliar virtual machines.
New languages.
New wallet standards.
New tooling.
Different deployment logic.
Completely separate ecosystems.
And then everyone wonders why developer activity slows down after the initial hype cycle.
EVM compatibility might sound boring compared to “autonomous AI economies,” but boring infrastructure decisions are usually the ones that determine adoption curves.
Solidity still dominates smart contract development. Ethereum tooling still dominates workflows. MetaMask is still the default wallet for huge parts of crypto. Hardhat, Foundry, Remix, ERC standards, Layer 2 deployment habits… developers already built years of muscle memory around these systems.
that part matters more than people think.
If an AI-focused blockchain forces developers to abandon all of that just to experiment with AI-native applications, many simply won’t bother. Especially experienced builders. Most teams optimize for speed and familiarity first.
OpenLedger seems to understand this tradeoff.
By staying EVM-compatible while building AI-focused infrastructure like Datanets, Proof of Attribution, model contribution tracking, and data monetization layers, the project lowers the mental overhead for Ethereum-native developers. A Solidity team can theoretically move faster because the environment already feels familiar.
That’s strategically important.
There’s also a deeper point here people sometimes miss: AI infrastructure itself needs composability. AI agents won’t operate in isolation forever. They’ll probably need access to liquidity, DeFi rails, identity systems, stablecoins, governance layers, and cross-chain settlement.
Ethereum already has those ecosystems.
So when OpenLedger talks about Ethereum settlement and modular infrastructure using OP Stack and EigenDA, it feels less like marketing language and more like an attempt to anchor AI systems into existing crypto liquidity networks instead of building another isolated chain nobody uses.
Still, EVM compatibility alone is definitely not enough.
And this is probably where the real test begins.
Almost every new chain now claims to be EVM-compatible. At this point it’s basically expected infrastructure. Developers hear “EVM support” the same way users hear “fast transactions.” It’s useful, but not differentiating by itself anymore.
The harder question is whether OpenLedger can actually build tooling developers want to use repeatedly.
Because AI infrastructure gets complicated very fast.
Model attribution sounds good conceptually, but attribution systems need to work cleanly under real network conditions. Data monetization sounds exciting too, but the incentive layers have to be practical, not theoretical. Developers need APIs, SDKs, indexing systems, debugging environments, reliable data availability, documentation that doesn’t feel half-finished, and onboarding that takes hours instead of weeks.
Otherwise even strong narratives fade.
I also think AI-native applications introduce a different type of pressure compared to traditional DeFi apps. AI workloads evolve rapidly. Models change constantly. Data pipelines shift. Throughput requirements become unpredictable. So scalability discussions around EigenDA and modular infrastructure probably matter more here than they do for simpler financial applications.
And honestly… crypto still hasn’t fully proven that decentralized AI infrastructure can outperform centralized alternatives in day-to-day usage.
That uncertainty is real.
OpenAI, Anthropic, and other centralized systems still dominate because they’re easy to use and deeply optimized. Decentralized AI projects need to compete on developer experience too, not just ideology. Builders won’t tolerate unnecessary friction simply because something is “more decentralized.”
That’s another reason why familiar Ethereum tooling helps.
Reducing friction is underrated strategy.
A developer who already understands Solidity, Ethereum wallets, ERC standards, L2 deployment flows, and existing EVM tooling can experiment with AI-focused products on OpenLedger without feeling like they entered a completely foreign environment. The barrier becomes lower psychologically and technically.
sometimes adoption is just momentum plus convenience.
I’m also curious how OpenLedger evolves its Datanets concept over time. If data contribution, attribution, and monetization actually become programmable primitives instead of abstract ideas, that could create interesting incentives for AI agents and autonomous applications. Especially if builders can compose those systems directly with existing Ethereum infrastructure.
That intersection feels more interesting to me than pure AI hype.
Not because it guarantees success. Nothing does in crypto. But because it aligns with how developers usually behave in practice.
They follow usable infrastructure.
They follow tooling.
They follow ecosystems that save time.
Narratives attract attention first. But workflows keep ecosystems alive.
And maybe that’s the bigger point here.
The future AI chains that survive probably won’t be the ones with the loudest branding. They’ll be the ones that quietly make developers feel productive from day one.
@OpenLedger seems aware of that reality. Whether it can fully execute on it is still an open question. But at least the direction makes sense.
In this market, practicality tends to outlast hype.
#OpenLedger $OPEN

GameFi cenu kustības vienmēr ir bijušas dīvainā veidā, kas stāsta tikai daļu no stāsta. Candlestick var parādīt nelielu pieaugumu vai strauju kritumu, bet tas reti atspoguļo to, kas patiesībā notiek zem virsmas — smalkas izmaiņas spēlētāju uzvedībā, neredzamās korekcijas spēļu sistēmās, klusa stimuli attīstība, ko spēlētāji jūt daudz pirms spēj to izskaidrot......
Tāpēc nesenais 2.2% pieaugums PIXEL pēdējo 24 stundu laikā neizskatās pēc virsraksta.... Tas vairāk izskatās pēc signāla. Kas ir kaut kas neliels virsmas līmenī, bet iespējams saistīts ar dziļām strukturālām izmaiņām, kas notiek iekšienē.

Sākumā es par to daudz nedomāju. @Pixels tas šķita kā vēl viens cikls... stādīt, gaidīt, novākt, atkārtot. Tīrs, paredzams. Tāds ritms, kur, ja ieguldi laiku, saņem kaut ko atpakaļ. Tā tam vajadzētu strādāt, vai ne?
Bet pēc kāda laika es sāku pamanīt nelielas plaisas. Nepietiekami lielas, lai to sauktu par sabojātu... vienkārši pietiekami, lai justos dīvaini. Dažās dienās es darīju to pašu rutīnu un viss bija kārtībā. Citās dienās šķita, ka tas pats darbs vienkārši... nesniedz to pašu rezultātu. Nekas acīmredzami nemainījās. Es nespēlēju sliktāk (vismaz tā es nedomāju). Tomēr rezultāti mainījās.