Google just opened up Gemini 1.5 Pro's 2-million-token context window to all developers, removing the waitlist. Both 1.5 Pro and Flash now also feature built-in Python code execution in an offline sandbox to tackle math and data reasoning, alongside Gemma 2 landing in Google AI Studio. Why it matters: developers can now feed entire codebases into Gemini while letting the model verify its own math by running code.
Anthropic just released Claude 3.5 Sonnet, leapfrogging its own flagship Opus and rival models on coding and reasoning benchmarks at twice the speed. It also introduced Artifacts, a new side-by-side workspace on Claude.ai that lets you view, edit, and run code or designs in real time. It is available for free on Claude.ai and via API.
DeepSeek just open-sourced DeepSeek Coder V2, an MoE-based code model supporting 338 languages and a 128K context window. It beat GPT-4 Turbo, Claude 3 Opus, and Gemini 1.5 Pro on major coding benchmarks like HumanEval and MBPP+, marking a rare open-source victory over leading proprietary models. Available under the MIT license in 16B and 236B sizes.
Mistral AI just dropped Codestral, its first-ever generative model fine-tuned specifically for coding, sporting 22B parameters and a 32k context window. Fluent in over 80 programming languages, it shines at fill-in-the-middle completions and repo-level tasks. Why it matters: Developers get an open-weight, low-latency engine directly inside their IDEs, though commercial deployments require a custom license.
Near is developing an AI agent designed to help non-coders build entire decentralized applications (DApps) simply by describing what they want. Co-founder Illia Polosukhin said the upcoming tool will generate smart contracts, frontends, and middleware end-to-end, pitching user-owned open-source AI as a counterweight to centralized tech giants.
Microsoft just added an AI-powered Advanced Paste tool to Windows 11 in PowerToys 0.81. The feature can convert clipboard text on the fly into Markdown, JSON, or plain text, and lets you prompt OpenAI to translate, summarize, or rewrite code across languages. The catch is that you need your own OpenAI API key and paid credits.
GitHub just launched Copilot Extensions, allowing developers to invoke third-party services like Docker, Sentry, and Azure directly inside Copilot Chat using natural language. Organizations can also build private extensions for internal APIs. Why it matters: developers can troubleshoot, manage databases, and deploy without leaving their IDE.
Anthropic just rolled out a prompt generator in the Anthropic Console to help developers create production-ready prompts. By simply describing what you want to achieve, Claude uses prompt engineering techniques like chain-of-thought reasoning to generate more effective and reliable prompts.
OpenAI and Stack Overflow just announced an API partnership to bridge technical knowledge with frontier LLMs. OpenAI will tap OverflowAPI to improve model performance and surface verified answers with attribution directly inside ChatGPT. In return, Stack Overflow will use OpenAI models to power its OverflowAI features, with the first integrations arriving in the first half of 2024.
GitHub just launched the technical preview of Copilot Workspace, a Copilot-native developer environment designed to build software from scratch in natural language. Starting from an issue or repo, agents brainstorm, plan, and test implementation while leaving developers in full control of every step. Why it matters: Coding is shifting rapidly from managing syntax to expressing intent.
AI coding startup Augment just emerged from stealth with a massive $252M funding round at a near-unicorn $977M valuation. Backed by former Google CEO Eric Schmidt and top venture firms, the company is founded by ex-Microsoft and Google veterans aiming to fix code quality and developer burnout. Why it matters: Coding assistants are a brutal, cash-burning market, but Augment now has the runway to take a real shot at Microsoft's GitHub Copilot.
Snowflake just launched Arctic LLM, an enterprise-focused open model licensed under Apache 2.0 for research and commercial use. Using a 480B-parameter MoE architecture that only activates 17B parameters at a time, it cost just $2 million to train and targets SQL generation and corporate coding workloads. Snowflake claims it outperforms rivals like Databricks’ DBRX on database tasks.
OpenAI just rolled out enterprise-grade controls for its API customers, introducing a Batch API that cuts prices by 50% for non-urgent tasks completed within 24 hours. The Assistants API also received a massive boost, expanding its file search limit by 500x up to 10,000 files alongside native streaming and project-level workspace controls.
Mistral AI just dropped its largest open-source model yet, releasing Mixtral 8x22B via an unceremonious torrent link on X. The sparse MoE model features 176 billion total parameters with 44 billion active per token and a 65,000-token context window. Released under a permissive Apache 2.0 license, it delivers massive compute efficiency to open-source builders without corporate guardrails.
OpenAI just dropped a majorly improved GPT-4 Turbo model in its API and began rolling it out to ChatGPT. GPT-4 Turbo with Vision is now generally available, finally letting developers use JSON mode and function calling directly on visual inputs. Why it matters: Multimodal workflows can now return reliable, structured data straight from images.
Google just released CodeGemma in collaboration with Hugging Face, bringing a family of open-access code specialists built on Gemma. Trained on an extra 500 billion tokens, the lineup includes an infilling-focused 2B model that runs on consumer hardware alongside 7B base and instruct variants. Why it matters: It offers developers fast, private code completion without needing massive cloud infrastructure.
Google just rolled out Gemini for Google Cloud, infusing its flagship AI across developer, security, and data tools. Duet AI is now Gemini Code Assist, powered by Gemini 1.5 Pro to analyze entire codebases across GitHub and GitLab, while new assistants streamline cloud architecture and BigQuery workflows.
OpenAI just expanded its fine-tuning API and custom models program, giving developers significantly more control over model adaptation. The update introduces epoch-based checkpoints, a comparative Playground UI for head-to-head testing, and assisted fine-tuning with OpenAI researchers. Why it matters: Developers and enterprises can now tailor frontier models to specialized domains with less guesswork and lower compute waste.
xAI just introduced Grok-1.5, boasting major upgrades in reasoning and math alongside a 128,000-token context window—a 16x jump over its predecessor. It scored 50.6% on the MATH benchmark and 74.1% on HumanEval. Early testers on 𝕏 will get access in the coming days.
Apple just announced that WWDC 2024 will run from June 10 to June 14. The main event will be iOS 18, expected to mark Apple's official push into generative AI with a smarter Siri, automated summarization, and AI code generation inside Xcode. Why it matters: Cupertino is finally ready to show its AI hand.
