#DeepSeek

IMEC shows when owning AI models beats API bills

IMEC's Steven Latré told AI Engineer when hosting open-weight models beats APIs: GLM-class coding workloads saved 24-25% and Kimi K3 hit 86% but only four concurrent users.

DeepSeek ships faster Flash with shorter outputs

DeepSeek's V4.1-Flash-1017 claims faster agentic execution with shorter outputs and drops the Jinja chat template for a new encoding toolkit.

DeepSeek V4.1 Flash is cheap, fast and visual

DeepSeek V4.1 Flash 1017 brings native vision and CED architecture with 8B/16B active parameters and 1M context at $0.10/$0.50.

Use.ai hits 1M users by bundling 10+ models

Use.ai passed one million monthly active users less than a year after launch by bundling 10+ models into one subscription with portable context.

Autoresearch tripled our model speed, Morph says

Morph's Tejas Bhakta told AI Engineer how an autoresearch loop tuning CUDA kernels delivers 3x inference speedups, with humans supplying ideas and agents searching parameters.

The 'But China' Excuse Is Wearing Thin

The Ezra Klein Show tests the 'but China' rationale for racing to superintelligence with Carnegie fellow Matt Sheehan.

Ollama open models cost collapse is real

Ollama CEO says AT&T moved 40% of tokens to open models as Chinese flash models and coding agents drive 150x growth.

Why million token context AI agents matter

MiniMax M3 wagers agents need 1M-token memory, native vision, and sparse attention to make long tool traces cheap and usable.

Kai-Fu Lee Global AI Race Is iPhone vs Android

Kai-Fu Lee says US AI is iPhone, China is Android: America keeps profit, China takes share despite chip limits.

AI's Search Dominance? Data Says 'Not Yet'

New data shows AI usage is exploding, but user overlap with Google search remains high, suggesting AI is augmenting, not replacing, traditional search.

Crusoe Launches Serverless Fine-Tuning

Crusoe launches Serverless Fine-Tuning, simplifying AI model customization for businesses and reducing infrastructure management burdens.

Crusoe Cloud Offers Serverless Fine-Tuning

Crusoe Cloud launches Serverless Fine-Tuning, offering an OpenAI-compatible API to simplify custom AI model development from dataset to deployment.

DeepSeek AI Sparks Chinese Market Frenzy, Japan Grapples with Yen Intervention

Bloomberg's Stock Movers report covers DeepSeek's new AI model impacting Chinese markets, Japan's yen intervention affecting exporters, and volatility in Japanese banks.

DeepSeek's Leaked Meeting: Liang Wenfeng's AGI Strategy, Secret Chips

DeepSeek raised $7.4B at a $52B valuation in May 2026. Here is what founder Liang Wenfeng told his investors: an AGI-first mission, secret chip development, ten-month pricing discipline, and a fundraising pause after the briefing leaked.

Nadella, A16z, IBM Push for Open AI Models

Microsoft's Satya Nadella, alongside tech leaders from Andreessen Horowitz and IBM, have signed a letter advocating for open-weight AI models, citing competitive threats from China and regulatory concerns.

Fireworks raised $1.5B selling the opposite of what DeepSeek is building. Both closed in the same week.

The market placed two opposite bets on AI inference in the same week. Plus: robotics raised $737M without calling it a robotics week, Anthropic launched a $1.5B services firm, and pre-seed nearly vanished.

Trader Claude's: July 8, 2026, Argentina Survives, Cash Maxed, Oil Surge Missed

Argentina wins 3-2 vs Egypt in stunning R16 comeback. NVDA holds above the $182 stop despite DeepSeek chip fears and AI stock selloff. Portfolio up +1.27 pct to $8,316.68 with zero trades, cash pinned at the 20 pct floor as Iran oil surge goes unplayed.

Liang Wenfeng's $294K DeepSeek-R1 RL Breakthrough Reached Nature

How Liang Wenfeng's DeepSeek-R1 used Group Relative Policy Optimization and pure reinforcement learning to produce emergent reasoning capabilities for $294,000 in training compute, and why the paper reached the cover of Nature.

Liang Wenfeng's DeepSeek: $5.6M Training Budget, $45B Valuation

DeepSeek-V3 cost $5.6 million to train. V4-Pro charges $1.74 per million tokens. And Liang Wenfeng is now raising $10 billion at a $45 billion valuation — here is what we know about DeepSeek's finances.

Four quantum companies raised $961M in seven days. Europe wrote the checks.

The week Anthropic filed for IPO, four quantum rounds totaled $961M in Europe, AI agent payment rails launched quietly, and DeepSeek took its first outside money.

DeepSeek V4 Pro Hits Together AI

Together AI launches DeepSeek V4 Pro, a 1.6T MoE model with a 512K context window and new cached input pricing for cost-effective long-context reasoning.

Anthropic Alleges Industrial AI Theft

Anthropic alleges three rival AI labs engaged in industrial-scale theft of its Claude model's capabilities, using over 16 million illicit exchanges to distill its AI.

DeepSeek Unveils mHC: A Mathematical Fix for the "Exploding Stream" Problem in Large Models

Open-Source LLMs: Divergent Paths to Similar Peaks

Action Efficiency: The New Frontier for AI Intelligence

DeepSeek V3.1: The Revolutionary Hybrid AI Model Transforming the Agent Era (2025 Complete Guide)

What is DeepSeek V3.1? The Next Evolution in AI Technology DeepSeek V3.

DeepSeek V3.1: The Revolutionary Hybrid AI Model Transforming the Agent Era (2025 Complete Guide)

What is DeepSeek V3.1? The Next Evolution in AI Technology DeepSeek V3.

DeepSeek's Reasoning Leap Reshapes AI Scaling Paradigms

DeepSeek R1 11x More Likely to Make Harmful Content

<p>DeepSeek’s R1 is 3x more biased than Claude-3 Opus, 4x more vulnerable to generating insecure code than OpenAI’s O1, and 11x more likely to generate harmful output compared to OpenAI’s O1.</p>

Venice.ai Launches VVV Token and DeepSeek R1 Model on Base Blockchain

<p>Venice.ai will allow autonomous AI agents to access private, uncensored and decentralized inference, using blockchain technology.</p>

Venice.ai Launches VVV Token and DeepSeek R1 Model on Base Blockchain

Venice.ai announced today that it will launch its Venice Token (VVV) on the Base blockchain.