# Compression İlgili Makaleler

HTX Haber Merkezi, kripto endüstrisindeki piyasa trendleri, proje güncellemeleri, teknoloji gelişmeleri ve düzenleyici politikaları kapsayan "Compression" hakkında en son makaleleri ve derinlemesine analizleri sunmaktadır.

Apple Re-invented Image Compression with AI: Same Quality, One-Third the File Size

Apple’s PICO: An AI-Powered Image Codec That Cuts File Size by Two-Thirds at Equal Perceived Quality In 2025, JPEG AI became the first international standard for learned image compression. However, it, like most codecs, still prioritizes mathematical metrics like PSNR over true perceptual quality—what the human eye finds pleasing. Apple researchers have introduced PICO (Perceptual Image Codec), a neural codec designed to optimize for human perception. It tackles key practical challenges: 1) Speed: A novel "one-shot context model" accelerates entropy encoding without sacrificing compression efficiency. 2) Artifacts: A dedicated TextFidelity loss preserves text clarity, and a TilingArtifact loss eliminates color seams between image tiles processed in parallel. 3) Control: It avoids the "hallucinations" common in GAN-based perceptual models. In a large-scale human evaluation (74,925 comparisons), PICO achieved the same perceived quality as standards like AV1, VVC, and JPEG AI while using only 30-43% of the bitrate. It also outperforms other learned perceptual codecs by 20-40%. Remarkably, it runs in 230ms (encode) and 150ms (decode) on an iPhone 17 Pro Max. While less efficient on synthetic graphics, PICO represents a significant shift from optimizing mathematical scores to directly targeting human visual experience, making high-quality perceptual compression practical for consumer devices. The work builds on expertise from WaveOne, whose team joined Apple and previously advanced neural video compression.

marsbit05/30 02:47

Apple Re-invented Image Compression with AI: Same Quality, One-Third the File Size

marsbit05/30 02:47

a16z: AI's 'Amnesia', Can Continuous Learning Cure It?

The article "a16z: AI's 'Amnesia' – Can Continual Learning Cure It?" explores the limitations of current large language models (LLMs), which, like the protagonist in the film *Memento*, are trapped in a perpetual present—unable to form new memories after training. While methods like in-context learning (ICL), retrieval-augmented generation (RAG), and external scaffolding (e.g., chat history, prompts) provide temporary solutions, they fail to enable true internalization of new knowledge. The authors argue that compression—the core of learning during training—is halted at deployment, preventing models from generalizing, discovering novel solutions (e.g., mathematical proofs), or handling adversarial scenarios. The piece introduces *continual learning* as a critical research direction to address this, categorizing approaches into three paths: 1. **Context**: Scaling external memory via longer context windows, multi-agent systems, and smarter retrieval. 2. **Modules**: Using pluggable adapters or external memory layers for specialization without full retraining. 3. **Weights**: Enabling parameter updates through sparse training, test-time training, meta-learning, distillation, and reinforcement learning from feedback. Challenges include catastrophic forgetting, safety risks, and auditability, but overcoming these could unlock models that learn iteratively from experience. The conclusion emphasizes that while context-based methods are effective, true breakthroughs require models to compress new information into weights post-deployment, moving from mere retrieval to genuine learning.

marsbit04/25 04:23

a16z: AI's 'Amnesia', Can Continuous Learning Cure It?

marsbit04/25 04:23

The Correct Way to Handle Claude Code Context: Avoid the Longer the Session, the 'Dumber' the Model

Anthropic has introduced a new /usage command for Claude Code, coinciding with its expanded 1 million token context window. This update addresses a key challenge: context rot, where model performance degrades as the context grows, due to attention dilution across excessive tokens. The article emphasizes strategic context management through five key actions: Continue, Rewind, Clear, Compact, and Subagents. It advises starting new sessions for new tasks but retaining relevant context for related work. The Rewind feature (double Esc or /rewind) is highlighted as a superior method for error correction, allowing users to revert and re-prompt from a specific point. Compaction (/compact) summarizes conversation history into a concise summary, though it can fail if the model misjudges future task relevance. In contrast, /clear requires manual summarization for a fresh start. Subagents are recommended for tasks generating disposable intermediate outputs, as they operate in isolated contexts and return only final results. The core recommendation is to proactively manage context—especially before model performance declines—to maintain high-quality outputs with Claude Code.

marsbit04/16 10:36

The Correct Way to Handle Claude Code Context: Avoid the Longer the Session, the 'Dumber' the Model

marsbit04/16 10:36

SUI compresses below $1.30, but $2.55 breakout is still possible – How?

SUI is experiencing significant compression near the $1.30 resistance level, with historical patterns suggesting potential for a strong upward breakout. Previous compressions led to rallies of up to 107%. The daily MACD shows bullish momentum, while the weekly chart forms a bullish pennant targeting a breakout above $2.55. A clean break could open a path toward the all-time high of $5.36. Supporting the bullish case, SUI recorded over $770 billion in cumulative trading volume, with strong recent weekly and daily activity outpacing major altcoins. High liquidity and growing participation indicate underlying demand despite current price consolidation.

ambcrypto03/02 07:03

SUI compresses below $1.30, but $2.55 breakout is still possible – How?

ambcrypto03/02 07:03

L1 Value Capture Shrinks Significantly, ETH, SOL, HYPE Struggle to Return to Price Peaks

The article "L1 Value Capture Shrinks Significantly: ETH, SOL, HYPE Struggle to Return to Price Peaks" argues that Layer-1 blockchains face a structural, not cyclical, problem: their ability to capture value through transaction fees is systematically eroded by innovation. Historically, periods of high demand (e.g., Bitcoin congestion, Ethereum's DeFi Summer, Solana's memecoin frenzy) create fee revenue peaks. However, these peaks inevitably stimulate the creation of cheaper alternatives that siphon away this income. The core finding is that open, permissionless networks cannot sustain high fee revenues; profitability is consistently competed away. **Key Examples:** * **Bitcoin:** Fee spikes from congestion (2017, 2021) were quickly mitigated by innovations like SegWit, batching, the Lightning Network, and wrapped BTC. The 2024-2025 bull run saw minimal fee growth despite a 3x price increase, with ETFs providing massive BTC exposure without on-chain fees. * **Ethereum:** The 2020-2021 fee boom from DeFi and NFTs was dismantled by competing L1s and, crucially, its own L2 scaling solutions. The Dencun upgrade (EIP-4844) drastically reduced data costs for L2s, causing Ethereum's L1 fee revenue to collapse by over 95% from its peak. * **Solana:** Its revenue relies heavily on MEV/tips from volatile memecoin trading. This income is now being compressed by private AMMs (which hide liquidity to prevent MEV) and platforms like Hyperliquid, which are moving the most profitable price discovery activity off-chain. **Impact on Token Valuation:** The market is shifting from valuing L1s based on "on-chain profit" to "asset narratives" and "structural capital flows." The analysis suggests: * **ETH:** Now resembles a low-yield infrastructure asset. Its fee compression is structural and ongoing. * **SOL:** While network activity may hit new highs, its matured fee-capturing mechanisms mean MEV revenue is unlikely to return to previous peaks, making a new all-time high price difficult. * **HYPE (Hyperliquid):** Currently benefits from high fees on its perp DEX. However, its fee model is under immense pressure to compress towards traditional finance (TradFi) rates (e.g., CME), threatening its projected high earnings and potentially its token price. * **BTC:** Its security model is unique and inverted. It relies almost entirely on block subsidies, not fees. Miner survival post-halving depends entirely on the USD price of BTC doubling to offset the 50% reduction in BTC-denominated rewards, making long-term security precariously tied to perpetual price appreciation.

marsbit02/26 08:45

L1 Value Capture Shrinks Significantly, ETH, SOL, HYPE Struggle to Return to Price Peaks

marsbit02/26 08:45

Arbitrum leads 2025 inflows, but ARB hesitates – What comes next in 2026?

Arbitrum led major blockchains in net inflows during 2025, reflecting a capital rotation toward scalable and reliable Layer 2 infrastructure, supported by strong fundamentals including $20B in total value locked and growing adoption like Robinhood's tokenized stocks. Despite steady user activity and revenue growth without incentive-driven surges, ARB's price remained subdued near $0.19, trading within a falling wedge with neutral momentum indicators. This suggests a tension between strong on-chain performance and short-term price hesitation heading into 2026.

ambcrypto12/29 03:02

Arbitrum leads 2025 inflows, but ARB hesitates – What comes next in 2026?

ambcrypto12/29 03:02

PEPE: Compression deepens as market awaits a decisive break

PEPE is at a critical juncture, compressing within a descending wedge pattern as leverage builds and traders await a decisive breakout. Key support lies at the $0.0000037–$0.0000039 demand zone, with a breakdown risking a fall toward $0.0000030. A breakout above wedge resistance near $0.0000050 could propel the price toward $0.0000063 and $0.0000079. Spot buyers are consistently absorbing sell pressure, providing a stabilizing base, while Open Interest has risen 7.6% to $222 million, indicating new leveraged positions. Long/short ratios show a fragile balance (52% vs. 48%), meaning a small move could trigger rapid, forceful expansion. The next break will likely be swift and decisive.

ambcrypto12/24 17:03

PEPE: Compression deepens as market awaits a decisive break

ambcrypto12/24 17:03

活动图片

Othersright-arrow

Regulatory Policyright-arrow