Matt White

@LinkedInHelp This automated locking and verification process that forces use of legal name contradicts the stated name policies, which allows for preferred professional names: linkedin.com/help/linkedin/…

@LinkedInHelp Hey, my account got locked because I'm trying to use a preferred name (official name change is in-progress), but it won't let me restore my account without using my dead name for it and won't let me contact support in any way.

The only game-changing behavior for me is that we finally have an OpenAI model that can code relatively competently. I'm not minimizing that - anything that can put downward price pressure on Anthropic is a win.

Top Performers in our internal benchmarks: GPT 5 - OpenAPI Spec Generation GPT 4.1 Mini - Text-to-SQL GPT 4.1 - Text-to-SQL (templated) Gemini 2.0 Flash Lite - Classification Gemini 2.0 Flash - Planning

I've had enough time working with the GPT-5 models (in Codex, Cursor, and Windsurf) and benchmarking them for performance in some of Alkemi's agentic tasks.

@lingvanex FYI, it's impossible to sign up as a US business self-service on your site. VAT Number is a required field. Can't provide what doesn't exist here.

There are still no other browsers that can really replicate how powerful Firefox containers can be, but Mozilla seems like a zombie. Even if Mozilla was the old Mozilla, too many sites don't support (or actively block) Firefox. I'm liking Vivaldi for now.

I held on as a @firefox user as long as I could. I don't want Chromium to be the only game in town - Google perpetually chipping away at privacy and ad blocking, but that's where lack of antitrust has brought us.

I've been trying out the agent mode for @cursor_ai more and more. o3-mini is supposed to be so good, but it still feels janky as the coding agent model compared to Claude Sonnet. Now if only @AnthropicAI could keep their servers up.

Trying out the AI app builder @Replit. So far it's a hot mess compared to co.dev

Updated chain of thought in OpenAI o3-mini for free and paid users, and in o3-mini-high for paid users.

I hear industry leaders like Sam Altman give lip service to democratizing AI, but I think this is what that actually looks like. It's great to see real competition heating up in the foundational LLM models. With any luck, LLM models is not going to be winner-take-all .

- make Gemini 2.0 Flash an attractive choice for any LLM application that requires quality and scalability. And the cost-optimized Gemini 2.0 Flash-Lite is somehow even cheaper.

Forget DeepSeek - Gemini 2.0 Flash is a wrecking ball to LLM pricing blog.google/technology/goo…

Introducing Citations. Our new API feature lets Claude ground its answers in sources you provide. Claude can then cite the specific sentences and passages that inform each response.

A Fully Local "Deep Researcher" w/ Deepseek-R1 @deepseek_ai-R1 is an open source, highly performant reasoning model. We used @ollama+LangGraph to build and test a local research assistant w/ Deepseek-R1-14b. Video: youtu.be/sGUjmyfof4Q Repo: github.com/langchain-ai/o…

@the_sukha_co Suddenly seeing the app crash due to an unhandled Spotify API error. Doesn't matter if I use Linux or Windows. Doesn't matter if I use Chrome, Firefox, or Brave. The app is unusable for me.