TPreço de Threshold Network Token

Investors Begin to “Change Tables” A notable shift is occurring in China's venture capital landscape. While the hard tech sector, especially AI, experiences a funding frenzy with record-breaking investment rounds, many investors are opting to leave traditional investment firms to join portfolio companies. This movement, termed "changing tables," marks a departure from past downturns where such moves were often a last resort. Currently, the trend is driven by proactive choice. Seasoned investors, including partners and managing directors, are being lured by attractive salaries, equity incentives, and core roles like VP, General Manager, or even co-founder at fast-growing startups in hot sectors like AI, embodied intelligence, and aerospace. This "going ashore" to companies is seen as a strategic career move. The shift is a two-way street. Booming fundraising and feverish investment activity have created intense competition among VCs for a narrow set of "star projects." Simultaneously, these high-flying tech companies, engaged in rapid, multi-round financing, urgently need talent with deep capital market expertise and institutional connections to manage their complex funding needs. Companies like Zhiyuan Robot, MoonDark AI, and MiniMax are actively recruiting from investment firms. For investors, the move also reflects concerns about the sustainability of the current market. Risks like valuation inversion between primary and secondary markets, high IPO破发 rates, increased regulatory scrutiny, and the uncertainty of eventual exits are prompting a reevaluation of the traditional VC path. While this transition offers new opportunities, it is not without challenges. Success requires adapting from evaluating companies to operating within one, focusing on execution over analysis. Nevertheless, this trend underscores the expanding career boundaries and fluidity within China's innovation ecosystem.

The founder and CEO of blockchain analytics firm Nansen, Alex Svanevik, predicts that Bitcoin is unlikely to ever fall below $60,000 again. He believes the crypto market is gradually shifting from a speculative model to using blockchain for real-world assets, with accumulation by large investors and ongoing demand supporting Bitcoin's price. Svanevik cites Bitcoin's role as a hedge against central bank money supply expansion as a key factor. Analysts from CryptoQuant note that the behavior of large holders around the $70,000 level (the realized price of accumulation addresses) will be a critical decision zone for the market. While acknowledging the recent sale of 1,638 BTC by MicroStrategy for corporate purposes, Svanevik also speculates that a Bitcoin price of $1 million could become the new normal if that level is reached. Separately, Svanevik stated that Solana is undervalued due to its misperception as primarily a memecoin platform, praising its development team and ecosystem. He views the tokenization of traditional assets as a major trend, positively assessing projects like Hyperliquid. Regarding the recently launched Robinhood Chain, he sees it as a potential competitor to Base, leveraging Robinhood's large user base, but questions the need for a separate Robinhood token.

AI-powered scrutiny of top scientific journals reveals that 99.2% of papers contain at least one error, challenging the perception of robust peer review. A recent study using AI agents to audit papers from ICML 2026 found that 58 out of 92 reviewed papers could not be fully reproduced. Reasons for failed replication include missing code, broken dependencies, and results inconsistent with claims. Separately, a GPT-5-based checker analyzed established AI conference papers, detecting an average of 4.7 objective errors per paper, with mathematical mistakes being most common. This trend suggests a growing "reproducibility crisis" as paper volume and complexity outpace traditional verification. However, it also presents an opportunity: researchers can now use AI to efficiently audit past literature, identify errors in foundational work, and publish corrections—a potentially fruitful new research avenue. In one case, AI even corrected century-old chemical data that had been accepted as fact. While AI tools significantly lower the cost of verification, they are not infallible (e.g., 83.2% precision rate in one system) and human oversight remains crucial. The era of AI-assisted verification may redefine the scientific process, where publication marks not an end, but the beginning of automated validation.

Harry Chun Tak Yeh, founder of Quantum Fintech Group, a notable participant in the Fantom ecosystem and crypto investor, has been found dead in Paraguay. His body was discovered at the foot of the elite Jade Park residential complex in Asunción. Police found his 30th-floor apartment open and ransacked. A 29-year-old Brazilian woman, identified as his partner and residing in another apartment in the same building, claimed no knowledge of the incident. Authorities are investigating multiple possibilities; Yeh's body, found naked and partially covered by a black plastic bag, has been sent for autopsy. Yeh was an early Bitcoin adopter, claimed to manage $2 billion in assets, and was involved with projects like LIF3 and L3 Reserve on Fantom.

The article discusses how AI Agents are transforming storage from a passive repository into an active component within AI inference and operation loops. As Agents perform continuous tasks involving models, memory, tools, and logs, data storage needs to evolve beyond simple block devices. The concept of "functional SSDs" is introduced, where capabilities like automatic encryption, compression, indexing, and memory management are embedded closer to the storage medium. This shift is driven by the need to handle Agent-specific data chains—including context, tool trajectories, and long-term memory—more efficiently. The piece analyzes current trends like AI SSDs from companies such as Phison (aiDAPTIV), Longsys (SPU+iSA), and Maxio, which are beginning to participate in the AI data path by managing model weights, KV cache, and prefetching. It further explores the future re-division of labor across the memory hierarchy: HBM for core compute, emerging High Bandwidth Flash (HBF) for read-intensive workloads, DRAM/CXL for mutable state, and functional SSDs for persistent, governed objects like Agent Memory. The conclusion is that storage will become integral to Agent capability, moving from just saving data to enabling next-step actions. The industry is poised to develop along three paths: functionalized SSDs, storage nodes tailored for Agents, and a re-architected, tiered memory system optimized for access patterns, security, and cost-per-token efficiency.