ONEЦена Harmony

In a recent interview, Sam Altman gave a rare and high praise, calling Alec Radford "perhaps the most important, yet least known, researcher in AI history." Widely regarded as the true father of GPT, Radford is the lead author of foundational papers including GPT-1, GPT-2, CLIP, and Whisper, and contributed significantly to GPT-3, DALL·E, Scaling Laws, and GPT-4. Despite this monumental impact, Radford remains highly private, holds no PhD, and rarely gives interviews. Altman credits OpenAI's rise to hiring the then-23-year-old in 2016. Radford's early experiments, like training a model on Amazon reviews, led to the discovery of the "unsupervised sentiment neuron," revealing that models can learn unintended capabilities from simple next-token prediction. His pivotal move was applying the Transformer architecture to language modeling, creating GPT-1 in 2018. This established the "scaling" direction—focusing on increasing model size, data, and compute—which became OpenAI's core strategy, leading to GPT-2, GPT-3, and beyond. Radford later applied the same principles beyond text. His early work on DCGAN laid groundwork for image generation. At OpenAI, he contributed to Image GPT, DALL·E, and CLIP, demonstrating that a simple, scalable training task (like matching images to text) could yield powerful, general capabilities. His work on Whisper applied this to robust speech recognition. Described by colleagues as a "once-in-a-generation genius" and exceptionally kind, Radford is known for his low profile. In late 2024, he left OpenAI for independent research. His latest project, "Talkie," is a 13-billion parameter language model trained *only* on texts published before 1931. This experiment tests if a model with no modern knowledge can quickly learn new skills (like basic Python) from few examples, probing the boundary between memorization and true learning. True to form, as the world catches up, Radford is likely already working on the next big question.

Harmony plans to roll back its blockchain to August 11th after an exploit created forged ONE tokens, discarding over 109,000 confirmed transactions. The network will revert to a specific block from that date and build a new chain, as selectively restoring transactions was deemed unsafe. The move follows reports of unauthorized ONE minting and distribution to exchanges, with investigators tracing most of the forged tokens. This rollback strategy places Harmony alongside networks like Ravencoin, which also sought to reverse confirmed blockchain activity after an exploit. Ravencoin recently faced a potential chain reorganization due to a consensus flaw.

Harmony plans to roll back its blockchain to a state from August 11th following an exploit that created counterfeit ONE tokens, and will invalidate over 109,000 transactions confirmed after that checkpoint. The Layer 1 network Harmony 1 stated that validators will revert to blocks recorded on August 11th at 11:25 UTC. New blocks will be created from subsequent heights using backup databases. The period to be invalidated includes 109,126 regular transactions and 315 staking transactions. Harmony noted that a selective rollback is not safe, as balances, smart contract states, nonces, and other parameters would differ in the new chain. The move comes after Harmony considered a rollback last week following reports of unauthorized ONE tokens being minted and sent to exchanges. The company stated investigators have traced nearly all counterfeit ONE to specific wallets or service providers and are working with exchanges, bridges, and law enforcement. This plan aligns Harmony with Ravencoin, another network seeking to reverse confirmed blockchain transactions after an exploit. Ravencoin recently faced a potential three-day blockchain reorganization due to an exploited consensus vulnerability, with mining pools controlling most of its hashrate starting to create a competing chain.

Compound, a pioneering decentralized lending protocol, has approved a record $52 million development budget and restructured its leadership to focus on institutional markets. This shift marks a move away from the retail yield-chasing that once drove DeFi. Its native token, $COMP, surged over 10% following the announcement. The protocol's Total Value Locked (TVL) has fallen about 90% from its 2021 peak to around $1.2 billion, lagging behind rivals like Aave. The budget, approved by the Compound DAO, allocates $28 million for operations and the new Compound V4 protocol, and $24 million for growth, including $8-10M specifically for institutional partnerships. V4 introduces a "hub-and-spoke" model for better capital efficiency and risk control. The new executive team includes veterans from Coinbase Custody, Near Foundation, and Maple Finance. Industry observers note this institutional pivot is a widespread survival strategy in crypto, as retail participation has dwindled and tokenized real-world assets grow.

A Cursor engineer's bold offer on X sparked a frenzy: send a screenshot of your canceled Claude subscription, get $200 in credits. This personal initiative by engineer Lauren (@poteto) promised a free month of Cursor's top-tier "Ultra" plan ($200/month, including Grok 4.6) to users switching from competitors like Claude Code. The post went viral, with users flooding her DMs. While some, like @notjazii, successfully received credits, many others faced radio silence or encountered barriers like X's premium membership requirement for messaging. Criticism arose over the chaotic, non-official handling, with some calling it a scam. The move is seen as an aggressive user-acquisition play following SpaceX's recent $60B acquisition of Cursor, directly targeting Claude Code's developer base. It highlights the intensifying subsidy wars in the AI coding tool space, putting a public price tag—$200—on developer loyalty.