# Daniil Sedov Russia, GMT+3 | [Telegram](https://t.me/gusarich) | [Github](https://github.com/gusarich) | [Email](mailto:gusarich@icloud.com) | [X](https://x.com/Gusarich) I work as an AI Research Engineer. I previously worked as a compiler engineer, and before that built and audited smart contracts. Posts with results of my research, as well as essays on other topics end up here. ## Pages - [LLM Tier List](https://gusarich.com/llm-tierlist) - [Codex Stats](https://gusarich.com/codex-stats) ## Posts - 27 January 2026 - [Things got too easy with AI](/blog/things-got-too-easy.md) - 2 January 2026 - [TON Vanity: 286,000x faster vanity addresses](/blog/ton-vanity.md) - 31 December 2025 - [AI in 2026](/blog/ai-in-2026.md) - 18 July 2025 - [Billions of Tokens Later: Scaling LLM Fuzzing in Practice](/blog/billions-of-tokens-later.md) - 26 March 2025 - [Documentation-Driven Compiler Fuzzing with Large Language Models](/blog/fuzzing-with-llms.md) [More posts ->](/blog.md) ## Blog posts - [My LLM Tier List](/blog/my-llm-tier-list.md) - [Things got too easy with AI](/blog/things-got-too-easy.md) - [I gave Codex its own Mac Mini](/blog/i-gave-codex-its-own-mac-mini.md) - [TON Vanity: 286,000x faster vanity addresses](/blog/ton-vanity.md) - [AI in 2026](/blog/ai-in-2026.md) - [What LLM to use today?](/blog/what-llm-to-use-today.md) - [There is nothing out-of-distribution](/blog/there-is-nothing-out-of-distribution.md) - [There is no singularity](/blog/there-is-no-singularity.md) - [Writing with AI](/blog/writing-with-ai.md) - [My impression of GPT-5](/blog/my-impression-of-gpt-5.md) - [The complexity threshold of AI](/blog/the-complexity-threshold-of-ai.md) - [Billions of Tokens Later: Scaling LLM Fuzzing in Practice](/blog/billions-of-tokens-later.md) - [Multitasking in 2025](/blog/multitasking-in-2025.md) - [Documentation-Driven Compiler Fuzzing with Large Language Models](/blog/fuzzing-with-llms.md) - [Measuring and Analyzing Entropy in Large Language Models](/blog/measuring-llm-entropy.md) See also: /llms-full.txt