#memory-architecture

[ follow ]
Artificial intelligence
fromInfoQ
2 days ago

Designing Memory for AI Agents: Inside Linkedin's Cognitive Memory Agent

LinkedIn's Cognitive Memory Agent enables context-aware AI systems that retain knowledge across interactions, enhancing personalization and continuity.
Science
fromFuturism
13 hours ago

Concern Grows That AI Is Damaging Users' Cognitive Abilities

Using ChatGPT for writing tasks may impair cognitive skills and creativity in students.
fromArs Technica
15 hours ago

AMD Ryzen 9 9950X3D2 Dual Edition review: Tons of cache for tons of dollars

What we didn't really find in our testing was evidence that the extra 64MB of L3 cache meaningfully improved performance beyond what the regular 9950X3D can already do.
Gadgets
#memory-chips
Apple
fromComputerworld
1 day ago

Is this where Apple Silicon will be in 5 years?

Apple is advancing chip technology with 2nm, 1.4nm, and sub-1nm processors, enhancing performance and energy efficiency across its hardware by 2030.
Node JS
fromDEV Community
3 days ago

I got tired of wiring the same caching stack every project, so I built LayerCache

LayerCache simplifies caching by stacking multiple layers and handling cache misses efficiently.
Mobile UX
fromGSMArena.com
3 days ago

Samsung to retire old LPDDR4 RAM and move production to LPDDR5

Samsung is phasing out LPDDR4 and LPDDR4X chips to focus on LPDDR5 production, enhancing performance and revenue.
Data science
fromInfoQ
1 week ago

Google's TurboQuant Compression May Support Faster Inference, Same Accuracy on Less Capable Hardware

TurboQuant compresses language models' Key-Value caches by up to 6x with near-zero accuracy loss, enabling efficient use of modest hardware.
Toronto startup
fromTNW | Tech
5 days ago

Musk is asking chip suppliers to move at 'light speed' for his Terafab AI

Musk's teams are actively seeking quotes for chipmaking equipment for the Terafab AI chip fabrication project.
#framework
Gadgets
fromEngadget
11 hours ago

Framework launches the Laptop 13 Pro with Intel's new Panther Lake chips

Framework's new 13 Pro laptop maintains modularity while introducing significant upgrades like a larger battery and redesigned chassis.
Gadgets
fromWIRED
11 hours ago

Framework Has a Better, More Take-Apart-Able Laptop

Framework announced the 13-inch Laptop 13 Pro, featuring improved battery life, a touchscreen, and modular design for easy upgrades.
Gadgets
fromEngadget
11 hours ago

Framework launches the Laptop 13 Pro with Intel's new Panther Lake chips

Framework's new 13 Pro laptop maintains modularity while introducing significant upgrades like a larger battery and redesigned chassis.
Gadgets
fromWIRED
11 hours ago

Framework Has a Better, More Take-Apart-Able Laptop

Framework announced the 13-inch Laptop 13 Pro, featuring improved battery life, a touchscreen, and modular design for easy upgrades.
#claude-code
Productivity
fromPerevillega
4 weeks ago

Building Agent Memory That Survives Between Sessions | Pere Villega

Memory in Claude Code sessions is a design problem requiring deliberate creation of context to avoid repetitive explanations.
#google
Tech industry
fromTNW | Artificial-Intelligence
2 days ago

Google in talks with Marvell Technology to build new AI inference chips alongside Broadcom TPU programme

Google is collaborating with Marvell Technology to develop new AI chips, enhancing its custom silicon supply chain for inference processing.
Tech industry
fromTNW | Artificial-Intelligence
2 days ago

Google in talks with Marvell Technology to build new AI inference chips alongside Broadcom TPU programme

Google is collaborating with Marvell Technology to develop new AI chips, enhancing its custom silicon supply chain for inference processing.
Artificial intelligence
fromnews.bitcoin.com
2 days ago

Nvidia Releases Nemotron 3 Super, a 120B Open AI Model Built for Agentic Workloads

Nvidia launched Nemotron 3 Super, a 120 billion parameter model that significantly reduces AI compute costs and increases throughput.
Science
fromNature
1 week ago

Brain-machine interface reveals the origin of a widely used neural signal

High gamma activity in the brain's cortex is primarily generated by synchronized neuronal inputs, impacting the interpretation of neuroscientific studies.
#ai
fromFuturism
3 days ago
Artificial intelligence

Study Finds AI Use Eats Away at Users' Confidence in Their Own Brains

from24/7 Wall St.
1 week ago
Tech industry

"Every Chip Is Getting Used Instantly" - Here's Why Google's AI Dominance May Be Unstoppable

Google's dominance in AI chip ownership positions it as the future leader in technology.
fromTNW | Corporates-Innovation
3 weeks ago
Data science

Google's TurboQuant compresses AI memory by 6x, rattles chip stocks

Google's TurboQuant algorithm significantly reduces memory usage for AI models, impacting memory stock prices due to lower physical memory needs.
Data science
fromTheregister
2 weeks ago

TurboQuant is a big deal, but it won't end the memory crunch

TurboQuant is an AI data compression technology that reduces memory usage for KV caches but may not significantly alleviate memory shortages.
Tech industry
from24/7 Wall St.
1 week ago

"Every Chip Is Getting Used Instantly" - Here's Why Google's AI Dominance May Be Unstoppable

Google's dominance in AI chip ownership positions it as the future leader in technology.
Software development
fromInfoWorld
5 days ago

The two-pass compiler is back - this time, it's fixing AI code generation

Multi-pass compilers revolutionized programming by separating analysis and optimization, a model that could enhance AI code generation.
Tech industry
fromThe Verge
3 days ago

The RAM shortage could last years

Manufacturers are expected to meet only 60% of DRAM demand by 2027, with shortages potentially lasting until 2030.
#intel
Gadgets
fromTheregister
4 days ago

Intel eases reliance on TSMC with Core Series 3 CPUs

Intel has introduced budget-oriented Core Series 3 processors manufactured in the US using a 2nm process, offering a solid upgrade for older systems.
Gadgets
fromGSMArena.com
4 days ago

Intel launches Core Series 3 processors for more affordable laptops

Intel unveils Core Series 3 processors for budget laptops, offering significant performance improvements and advanced features for consumers and small businesses.
Gadgets
fromEngadget
5 days ago

Intel launches new Core Series 3 chips for mainstream laptops

Intel's new Core Series 3 chips offer significant performance improvements and exceptional battery life for mainstream laptops.
Gadgets
fromTheregister
4 days ago

Intel eases reliance on TSMC with Core Series 3 CPUs

Intel has introduced budget-oriented Core Series 3 processors manufactured in the US using a 2nm process, offering a solid upgrade for older systems.
Gadgets
fromGSMArena.com
4 days ago

Intel launches Core Series 3 processors for more affordable laptops

Intel unveils Core Series 3 processors for budget laptops, offering significant performance improvements and advanced features for consumers and small businesses.
Gadgets
fromEngadget
5 days ago

Intel launches new Core Series 3 chips for mainstream laptops

Intel's new Core Series 3 chips offer significant performance improvements and exceptional battery life for mainstream laptops.
Tech industry
fromEngadget
4 days ago

A lot of you panic-bought PCs to avoid RAMaggedon 2026

AI-driven demand for memory and storage is causing pre-emptive computer purchases, leading to a 3.2% increase in global PC shipments in Q1 2026.
Software development
fromInfoQ
2 weeks ago

When Every Bit Counts: How Valkey Rebuilt Its Hashtable for Modern Hardware

Redis clones offer opportunities for optimization and learning, but often lack full implementation and reliability essential for caching.
Science
fromNature
2 weeks ago

Breakthrough computer chip tech could help meet 'monumental demand' driven by AI

A new light source enables the creation of 8 nm wide structures on silicon wafers, increasing transistor density for advanced computer chips.
Artificial intelligence
fromEngadget
6 days ago

There's yet another study about how bad AI is for our brains

AI assistance improves immediate performance but creates dependency, leading to decreased persistence and independent performance when the technology is removed.
DevOps
fromInfoWorld
4 weeks ago

An architecture for engineering AI context

AI systems must intelligently manage context to ensure accuracy and reliability in real applications.
Environment
fromFast Company
1 month ago

This is why helping people remember is the best strategy

Radical leadership involves helping people remember what is essential in a world obsessed with constant growth and productivity.
Tech industry
from24/7 Wall St.
2 weeks ago

Micron's AI Memory Boom Is Real-And Analysts Are Still Playing Catch Up

Micron shares remain volatile amid questions about memory demand due to AI efficiency gains and potential supply shortages into the 2030s.
Roam Research
fromNature
1 month ago

'RAMmageddon' hits labs: AI-driven memory shortage is impacting science

RAM shortage driven by AI demand is causing prices to triple, affecting scientists and researchers who need computing resources through 2027.
Artificial intelligence
fromMedium
2 weeks ago

Hindsight: The Future of AI Agent Memory Beyond Vector Databases

Hindsight introduces a new AI memory system that enables learning from experiences rather than just recalling past information.
#biological-computing
Science
fromFuturism
1 month ago

Staff at New Data Center Powered by Human Brain Cells Need to Swap Out Cerebrospinal Fluid Every Day

Cortical Labs' biological computers require constant replenishment of cerebrospinal fluid and have unique operational needs compared to traditional data centers.
Science
fromFuturism
1 month ago

New Data Centers Will Be Powered by Human Brain Cells

Cortical Labs is building biological data centers using living human neurons as computing units, consuming far less power than traditional AI processors.
Science
fromFuturism
1 month ago

Staff at New Data Center Powered by Human Brain Cells Need to Swap Out Cerebrospinal Fluid Every Day

Cortical Labs' biological computers require constant replenishment of cerebrospinal fluid and have unique operational needs compared to traditional data centers.
Science
fromFuturism
1 month ago

New Data Centers Will Be Powered by Human Brain Cells

Cortical Labs is building biological data centers using living human neurons as computing units, consuming far less power than traditional AI processors.
fromZDNET
2 weeks ago

Is increasing VRAM finally worth it? I ran the numbers on my Windows 11 PC

Rising RAM prices have made upgrading your PC more expensive. Virtual RAM is a less expensive way of boosting an older computer's performance, but it has limited use cases because it can't match the speed of physical RAM.
Gadgets
#arm
Tech industry
fromWIRED
4 weeks ago

Arm Is Now Making Its Own Chips

Arm is producing its own semiconductors, marking a shift from licensing to manufacturing in response to AI demand.
Tech industry
fromWIRED
4 weeks ago

Arm Is Now Making Its Own Chips

Arm is producing its own semiconductors, marking a shift from licensing to manufacturing in response to AI demand.
#amd
Gadgets
fromTheregister
3 weeks ago

AMD doubles up on V-Cache with 9950X3D2 Dual Edition

AMD's Ryzen 9 9950X3D2 Dual Edition CPU enhances gaming and production performance with 16 cores and 192 MB L3 cache.
Gadgets
fromEngadget
3 weeks ago

AMD's Ryzen 9950X3D2 chip features an incredible 208MB of on-chip cache

AMD's Ryzen 9950X3D2 features 208MB of cache and is designed for gaming and creative workloads.
Gadgets
fromTheregister
3 weeks ago

AMD doubles up on V-Cache with 9950X3D2 Dual Edition

AMD's Ryzen 9 9950X3D2 Dual Edition CPU enhances gaming and production performance with 16 cores and 192 MB L3 cache.
Gadgets
fromEngadget
3 weeks ago

AMD's Ryzen 9950X3D2 chip features an incredible 208MB of on-chip cache

AMD's Ryzen 9950X3D2 features 208MB of cache and is designed for gaming and creative workloads.
Tech industry
fromThe Verge
4 weeks ago

Arm's first CPU ever will plug into Meta's AI datacenters later this year

Arm AGI CPU features up to 136 cores and claims double the performance per watt compared to x86 chips.
Miscellaneous
fromInfoQ
1 month ago

OpenAI Codex-Spark Achieves Ultra-Fast Coding Speeds on Cerebras Hardware

OpenAI deployed GPT-5.3-Codex-Spark on Cerebras wafer-scale chips, achieving 1,000 tokens per second for real-time interactive coding with 15× faster performance than earlier versions.
Artificial intelligence
fromMedium
4 weeks ago

Less Compute, More Impact: How Model Quantization Fuels the Next Wave of Agentic AI

Model quantization and architectural optimization can outperform larger models, challenging the belief that more GPUs equal greater intelligence.
fromTheregister
1 month ago

RAM is getting expensive, so squeeze the most from it

Both work with Linux's existing swapping mechanism. Swapping (called paging in Windows) is a way for the kernel to handle running low on available RAM. It chooses pages of memory that aren't in use right now and copies them to disk, then those blocks can be marked as free and reused for something else.
Software development
Tech industry
fromTheregister
1 month ago

Your next car night need 300GB of RAM, and so will robots

Micron Technology projects sustained growth driven by AI demand, with autonomous vehicles and humanoid robots requiring 300+ gigabytes of DRAM, while the company tripled revenue year-over-year to $23.86 billion in Q2 2026.
Tech industry
fromTechzine Global
1 month ago

Samsung and AMD strengthen collaboration on HBM4 for AI chips

Samsung and AMD expand collaboration to supply HBM4 memory for MI455X accelerators, DDR5 for EPYC processors, and explore foundry partnership for next-generation products.
fromGSMArena.com
1 month ago

AMD buys HBM4 from Samsung for AI data centers

Samsung and AMD share a commitment to advancing AI computing, and this agreement reflects the growing scope of our collaboration. From industry-leading HBM4 and next-generation memory architectures to cutting-edge foundry and advanced packaging, Samsung is uniquely positioned to deliver unrivaled turnkey capabilities that support AMD's evolving AI roadmap.
Tech industry
Artificial intelligence
fromTechCrunch
1 month ago

Niv-AI exits stealth to wring more power performance out of GPUs | TechCrunch

AI data centers waste significant power due to GPU demand surges, forcing operators to throttle performance by up to 30%, prompting startups like Niv-AI to develop precision power management solutions.
#ai-chips
fromTechzine Global
1 month ago
Artificial intelligence

Meta shifts to AI inference with its future chips

Meta developed four MTIA AI chip generations in under two years, shifting focus from content ranking to GenAI inference, with production and deployment planned through 2027.
Tech industry
fromWIRED
1 month ago

Meta Developed Four New Chips to Power Its AI and Recommendation Systems

Meta developed four new AI chips (MTIA 300, 400, 450, 500) for powering generative AI and content ranking, with one in production and three shipping between 2027.
Tech industry
fromWIRED
1 month ago

Meta Developed Four New Chips to Power Its AI and Recommendation Systems

Meta developed four new AI chips (MTIA 300, 400, 450, 500) for powering generative AI and content ranking, with one in production and three shipping between 2027.
Gadgets
fromTheregister
2 months ago

Cache is king and DIMMS are bling as memory prices soar

Second-hand server memory prices have surged roughly 700%, creating potential resale windfalls for home-lab owners holding large quantities of RAM and enterprise SSDs.
Gadgets
fromFast Company
1 month ago

Here's every cool tech thing the AI RAM crunch is ruining

AI data-center demand is diverting DRAM production to server RAM, causing consumer RAM shortages, device delays, and steep price increases through 2028.
Artificial intelligence
fromTechCrunch
2 months ago

Running AI models is turning into a memory game | TechCrunch

Rising DRAM prices and sophisticated prompt-caching orchestration make memory management a critical cost and performance factor for large-scale AI deployments.
Gadgets
fromTheregister
2 months ago

Micron's first PCIe 6.0 SSDs are all about AI

Micron's PCIe 6.0 9650 SSDs deliver up to 28 GB/s transfer rates and are targeted at datacenter AI storage rather than consumer PCs.
Artificial intelligence
fromLogRocket Blog
2 months ago

Building AI apps that remember: Mem0 vs Supermemory - LogRocket Blog

Long-term memory is essential for LLM applications to be stateful, preserving user context and preferences across sessions for efficient, connected experiences.
fromTheregister
2 months ago

Positron opts for laptop RAM over HBM to take on Nvidia

On paper, Positron's next-gen Asimov accelerators, no doubt named for the beloved science fiction author, don't look like much of a match for Nvidia's Rubin GPUs. Yet, the Arm-backed AI startup boasts its inference chip will churn out five times as many tokens per dollar while using one-fifth the power of Nvidia's latest accelerators to do it. Those are certainly some bold claims, which the company contends are possible because the chip was designed to support large-scale inference workloads.
Artificial intelligence
fromTheregister
2 months ago

Intel greets memory apocalypse with Xeon workstation CPUs

The Xeon 600 lineup spans the gamut between 12 and 86 performance cores (no cut-down efficiency cores here), with support for between four and eight channels of DDR5 and 80 to 128 lanes of PCIe 5.0 connectivity. Compared to its aging W-3500-series chips, Intel is claiming a 9 percent uplift in single threaded workloads and up to 61 percent higher performance in multithreaded jobs, thanks in no small part to an additional 22 processor cores this generation.
Tech industry
Tech industry
fromTechzine Global
2 months ago

Samsung nears Nvidia approval for HBM4 memory

Samsung is nearing Nvidia approval and February mass production for HBM4 AI memory, narrowing the gap with SK Hynix amid an AI-driven memory shortage.
fromTechCrunch
2 months ago

Quadric rides the shift from cloud AI to on-device inference - and it's paying off | TechCrunch

The company, which is based in San Francisco and has an office in Pune, India, is targeting up to $35 million this year as it builds a royalty-driven on-device AI business. That growth has buoyed the company, which now has post-money valuation of between $270 million and $300 million, up from around $100 million in its 2022 Series B, Kheterpal said.
Artificial intelligence
fromLogRocket Blog
2 months ago

Building AI apps that remember: Mem0 vs Supermemory - LogRocket Blog

Large Language Models (LLMs) enable fluent, natural conversations, but most applications built on top of them remain fundamentally stateless. Each interaction starts from scratch, with no durable understanding of the user beyond the current prompt. This becomes a problem quickly. A customer support bot that forgets past orders or a personal assistant that repeatedly asks for preferences delivers an experience that feels disconnected and inefficient.
Artificial intelligence
Artificial intelligence
fromTheregister
2 months ago

How agentic AI strains modern memory hierarchies

Agentic AI shifts the system bottleneck from raw compute to memory: prolonged KV cache residency demands greater capacity, bandwidth, and fast hierarchical memory switching.
Artificial intelligence
fromTheregister
2 months ago

OpenAI unveils first model running on Cerebras silicon

OpenAI launched GPT-5.3-Codex-Spark on Cerebras' SRAM-packed accelerators to deliver over 1,000 tokens per second with a 128K token context window.
Artificial intelligence
fromInfoWorld
2 months ago

What is context engineering? And why it's the new AI architecture

Context engineering designs and manages the information, tools, and constraints an LLM receives, enabling scalable, high-signal inputs and improved model outcomes.
Artificial intelligence
fromComputerworld
2 months ago

Microsoft launches its second generation AI inference chip, Maia 200

Maia 200 is a high-performance, energy-efficient inference accelerator optimized for large reasoning models, delivering superior FP4/FP8 throughput and memory compared with rival cloud accelerators.
Artificial intelligence
fromInfoWorld
1 month ago

Why AI requires rethinking the storage-compute divide

AI workloads require continuous processing of unstructured multimodal data, causing redundant data movement and transformation that wastes infrastructure costs and data scientist time.
[ Load more ]