Bigachu retweeted
Apple's in talks with @PrismML, who is shrinking a large open model to run on iPhone.
Notice what's missing: the researchers who built the model. No mention, no cut, no say in how it's used. That's not an edge case, it's what "open weights" currently means by default.
OML fixes the missing piece: fingerprinting empowers researchers to prove ownership, and makes unauthorized use impossible to hide.
Bigachu retweeted
$NVDA $TSM "Next-Gen Rosa CPU May Adopt TSMC A16 with Back-Side Power Delivery, Eyes 2028 Launch"
"TSMC’s A16 offers up to 1.10× higher chip density than N2P. Its biggest technological breakthrough is the introduction of Super Power Rail (SPR), a back-side power delivery technology that moves the power network to the back of the wafer. This improves power delivery efficiency, reduces IR drop, and frees up front-side routing space for signal interconnects.”
“Rosa CPU for agentic AI will feature its in-house Rigel cores based on the Arm v9.2 instruction set. Compared with the Olympus cores used in the Vera CPU, Rigel is designed to deliver higher single-core performance within the same die area. Key enhancements include improved instruction delivery, a larger L2 cache, and more efficient memory management.”
“Although NVIDIA has not disclosed the exact core count or IPC improvements for the Rosa CPU, it is expected to feature 128 or more Rigel cores, with IPC performance potentially improving by more than 50%.”
Bigachu retweeted
Meet Hy3 on ModelScope!
295B total / 21B active MoE, built for agentic workflows with 256K context and an FP8 variant ready for deployment. 🚀 License: Apache 2.0.
🤖modelscope.ai/collections/Te…
🏆 Human eval: 270 experts compared outputs in blind real-world workflow tests. Hy3 scores 2.67/4, ahead of GLM-5.1 at 2.51/4, with the clearest gains in frontend development, CI/CD, and data & storage.
🛠️ Agent reliability: stronger tool calling, formatting, and error recovery, with SWE-Bench Verified variance within 4% across major scaffolds.
🧠 Multi-turn gains: issue rate drops from 17.4% to 7.9%, and MRCR rises from 42.9% to 75.1%.
⚙️ Model lineup: Hy3 BF16 instruct model plus Hy3-FP8 quantized instruct model, with vLLM and SGLang deployment support.
Bigachu retweeted
Even Zuckerberg is frustrated with AI agents not moving fast enough.
At an internal Meta meeting, Mark Zuckerberg reportedly told staff that AI agent development hasn't progressed as quickly as he'd hoped.
This is the same guy who bet the entire company on AI. Meta is spending billions. They have Llama. They have internal infra most startups can only dream of.
And even HE is saying - it's not fast enough.
That tells you something real about where AI agents actually are right now vs the hype. Everyone is talking about autonomous agents taking over workflows. The demos look incredible. But production-grade, reliable agents that actually work at scale? Still a hard problem.
The gap between "impressive demo" and "deployed at Meta scale" is massive.
So next time someone tells you agents are already here and your job is gone — maybe wait a bit. 😅
Even Zuck is waiting.
Bigachu retweeted
How did IT services become so big in India?
Not because of our great skills, but because the West wanted to replicate its blue-collar experiments on white-collar work.
From the 1980s, outsourcing blue-collar work to China helped big global manufacturers cut their wage bills sharply and increase their profit margins.
But their white-collar salary bills ballooned, because they needed many more people to manage global supply chains.
The internet opened the door to outsourcing white-collar work too — by handing it over to remote workers sitting in the Third World.
Indian IT companies — which already had some experience with managing remote systems — were best placed to capture this market. They had access to a wide pool of English-educated computer science graduates, who could do the same work at a fraction of the cost.
Now, AI agents are turning out to be even cheaper than developers sitting in Bengaluru. At the very least, AI companies have the funding to subsidise their agents and lock in corporate clients for years.
The same logic that built Indian IT — cheaper labour — is now dismantling its business model.
Life has come full circle for India's IT giants.
79% of revenue from transaction charges, 60% from a single weekly options product. that's not an exchange, that's a Nifty 50 derivatives shop with a clearing license attached.
We did a detailed breakdown of NSE DRHP on @zerodhamarkets Daily Brief.
Read: thedailybrief.zerodha.com/p/…
Watch: youtube.com/watch?v=KJUOeXLo…
Bigachu retweeted
Markets are closed today.
Volatility is never easy.
Bitcoin keeps working.
So do we.
Thank you for your support.
Bigachu retweeted
Absolutely incredible:
Elon musk is now worth more than the entire market cap of Bitcoin.
You're SO early.
Bigachu retweeted
AI agents combine LLMs with tools, memory, and workflows so they can fetch data, call APIs, and take actions.
And in this tutorial, Manoj teaches you how to build your own AI agent using LangGraph.
You'll learn about the tools, agent state, and the agent graph - and how to wire it all together.
freecodecamp.org/news/how-to…
Bigachu retweeted
if you haven't looked at what @SentientAGI has going on with their current eval cycle, worth a glance. entries this round are running higher than previous cycles, which tells you something about where community attention is sitting right now.
xitter.cf/SentientAGI/status/206…
Bigachu retweeted
$SUI - One thing I think market still underestimates about Sui is how much activity has already been built beyond the "fast L1" narrative.
As of June 2026:
- ~719K monthly active addresses.
- 39M+ token holders.
- ~$17.6M cumulative fees generated.
- ~$358M stablecoin supply on-chain.
More importantly, growth is no longer coming from a single sector.
Sui has gradually expanded across:
- DeFi (Cetus, Navi, Scallop, Suilend, DeepBook).
- Stablecoins & payments.
- BTCFi products.
- Gaming & consumer applications.
- RWA initiatives with institutional partners.
Looking at current data, $Sui has clearly made meaningful progress across user growth, ecosystem expansion and institutional partnerships over the past two years.
However, growth alone doesn't automatically translate into long-term value.
Revenue generation remains relatively small, tokenomics still face significant dilution pressure, and sustainable competitive advantages are not yet fully established.
For now, I see #Sui as a promising ecosystem with strong technology and growing adoption, but still a long way from proving that activity can consistently translate into durable economic value.
Framework Rating: 2.7/5 ⭐
Technology (3.5/5) | Ecosystem (3/5) | User Activity (4/5) | Partnerships (3/5) | Revenue (1.5/5) | Token Capture (1.5/5)
A growing ecosystem, strong product, and one of the more interesting L1s to watch this cycle.
$VIRTUAL is one of the best examples of how powerful a narrative can be in crypto.
The thesis is simple: instead of building a single AI product, @virtuals_io aims to become the marketplace where AI agents are created, owned, and monetized. If AI agents become a major category in crypto, #Virtuals could be one of the biggest beneficiaries.
The ecosystem has grown rapidly:
> 46,500+ AI projects launched.
> $481M Agent GDP.
> $13.8B cumulative trading volume.
> 2.38M jobs completed.
These numbers are impressive, but the key question is whether activity is translating into sustainable economic value.
For me, the most important metrics are not the number of agents launched, but:
> Are users returning and actively using these agents?
> Are agents generating recurring revenue?
> Does the value created ultimately flow back to the protocol and token?
This is where I think the market may be getting ahead of reality. Ecosystem activity is growing, but the business fundamentals still look less mature than the current valuation suggests. Much of the thesis depends on AI agents evolving from speculative assets into products people genuinely use.
The bull case is clear:
> AI agents become a major crypto category.
> Virtuals remains the leading launch platform.
> Network effects and economic activity continue to strengthen.
The bear case is equally straightforward:
> Most agents fail to achieve meaningful PMF.
> Revenue growth remains limited.
> Narrative value continues to outpace economic value.
At this stage, I view $VIRTUAL as a promising experiment in AI-native economies rather than a proven business. The opportunity is significant, but there is still a lot that needs to be proven before the fundamentals fully support the story.
My Framework Rating for $VIRTUAL: 2.4/5 ⭐
Narrative (4/5) | Brand (3.5/5) | Ecosystem (2.5/5) | PMF (2/5) | Network Effects (2/5) | Revenue (1.5/5) | Token Capture (1.5/5)
The narrative remains stronger than the economics. Until revenue, retention, and token value capture improve meaningfully, I still see $VIRTUAL as a promising experiment rather than a mature business.
