<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><description>Member of Technical Staff@Cantina Labs (NYC);&#xA;CS PhD @CUBoulder;&#xA;Posts are my own;&#xA;Website: https://akkikiki.github.io/&#xA;Substack: http://substack.com/@akkikiki&#xA;Lived: 🇹🇭🇯🇵🇫🇷🇺🇸&#xA;Posts: JA/EN</description><link>https://bsky.app/profile/akkikiki.bsky.social</link><title>@akkikiki.bsky.social - Yoshinari Fujinuma</title><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3mf4pfugsdk2l</link><description>If you&#39;re using Claude Code, I highly recommend Claude-Code-tools for enhancing productivity, especially the voice plugin among them, which responds with TTS.&#xA;&#xA;And if you get stuck on installation, just ask Claude Code. It should take less than 10 minutes.&#xA;&#xA;https://github.com/pchalasani/claude-code-tools</description><pubDate>18 Feb 2026 08:54 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3mf4pfugsdk2l</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3m5sctobmkc2v</link><description>Does anybody know the list of social events happening at NeurIPS @ San Diego? Only found luma.com/NeurIPS so far&#xA;https://luma.com/NeurIPS</description><pubDate>17 Nov 2025 03:47 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3m5sctobmkc2v</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lo2r7bglxk24</link><description>I played around with the mlx and 4-bit version of Qwen3-30B-A3B locally on an Apple M4 Max chip, with Japanese and English. This is amazing. It seems feasible to run it locally for tasks other than long-horizon or complex ones.</description><pubDate>30 Apr 2025 21:42 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lo2r7bglxk24</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lgbmg4md6c2k</link><description>&#xA;I&#39;ve just played around with DeepSeek-R1 and wow, such a long thoughts for a simple question &#34;What is the square root of 16?&#34; 😀</description><pubDate>21 Jan 2025 19:43 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lgbmg4md6c2k</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lchfkvfwmk2k</link><description>My LinkedIn feed is full of AWS re:Invent posts (since I work at AWS, and many colleagues share about it), Twitter/X is a mixture of everything, and Bluesky posts are mostly academic. Welcome to the filter bubbles!</description><pubDate>04 Dec 2024 05:16 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lchfkvfwmk2k</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lc2tdadcyk2a</link><description>I was taking a stab at the responding to author discussions for ARR for the last few days, but some common issues I see is that submitted drafts are pretty exaggerating how good the results are.</description><pubDate>29 Nov 2024 05:17 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lc2tdadcyk2a</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lbt3tbi74c2x</link><description>Where is the open-weight model that we can try it out? 😀 &#xA;https://blogs.nvidia.com/blog/fugatto-gen-ai-sound-model/</description><pubDate>26 Nov 2024 03:28 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lbt3tbi74c2x</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lbsurceptk2x</link><description>I like typeset.io/pdf-to-video &#39;s pdf-to-video feature for getting a quick overview of the paper. Looking forward to having more fine-grained video (or even customizable controlled generation of pdf video summary) version of it 😀</description><pubDate>26 Nov 2024 01:22 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lbsurceptk2x</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lbleu5fres2i</link><description>Here&#39;s the starter pack for AI/ML/NLP conferences that I was able to find as of now. I couldn&#39;t remove myself from the starter pack so feel free to unfollow me after hitting the &#34;follow all&#34; button 🙂 go.bsky.app/9QQXJ1u&#xA;&#xA;[contains quote post or other embedded content]</description><pubDate>23 Nov 2024 01:49 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lbleu5fres2i</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lbkqg77ar22s</link><description>Someone should really treat me some coffee for asking me to assign &amp; finish the emergency review within a day 😀</description><pubDate>22 Nov 2024 19:43 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lbkqg77ar22s</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lbggwyd2ak2a</link><description>Today&#39;s paper reading: Interesting, quantization can reverse unlearning up to 83%  arxiv.org/abs/2410.16454&#xA;https://arxiv.org/abs/2410.16454</description><pubDate>21 Nov 2024 02:43 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lbggwyd2ak2a</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3lbeaaimq7g25</link><description>Today&#39;s paper reading: arxiv.org/abs/2306.13649&#xA;https://arxiv.org/abs/2306.13649</description><pubDate>20 Nov 2024 05:37 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3lbeaaimq7g25</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kh7a2vcwtw2e</link><description>Too many people... @Shibuya station, Tokyo, Japan</description><pubDate>23 Dec 2023 08:56 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kh7a2vcwtw2e</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kfysq5lqw62b</link><description>Wow, more than 2000 papers were accepted in total for EMNLP</description><pubDate>08 Dec 2023 02:16 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kfysq5lqw62b</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kfuek24ps22m</link><description>Heading to #EMNLP ! Co-authored papers 👇&#xA;[1] 12/9 11am in-person poster by Sharon Levy title &#34;Comparing Biases and the Impact of Multilingual Training across Multiple Languages&#34; &#xA;&#xA;[2] 12/8 2pm virtual poster titled &#34;A Multi-Modal Multilingual Benchmark for Document Image Classification&#34;</description><pubDate>06 Dec 2023 07:51 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kfuek24ps22m</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kckfh5zqyv2x</link><description>Resolving Latex errors for uploading to arxiv...</description><pubDate>25 Oct 2023 04:26 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kckfh5zqyv2x</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kchnxdmfgt2r</link><description>GPT-4V just fixed my circuit breaker (where I had been struggling for 10+ mins at midnight)</description><pubDate>24 Oct 2023 02:20 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kchnxdmfgt2r</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kca5npbfze2o</link><description>Done finishing up the EMNLP findings camera ready</description><pubDate>21 Oct 2023 02:39 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kca5npbfze2o</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kc2y6tvnjk2p</link><description>I finally read the attention sink paper [1] and the HF blog article [2]. Seems like another interesting data point that the models we usually interact with strongly attend to the first few tokens...&#xA;[1] arxiv.org/abs/2309.17453&#xA;[2] https://huggingface.co/blog/tomaarsen/attention-sinks</description><pubDate>19 Oct 2023 01:18 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kc2y6tvnjk2p</guid></item><item><link>https://bsky.app/profile/akkikiki.bsky.social/post/3kbqnek5ffv2f</link><description>Preparing an EMNLP camera ready version of our accepted paper ✍️✍️✍️</description><pubDate>14 Oct 2023 22:38 +0000</pubDate><guid isPermaLink="false">at://did:plc:2v4ja57f2yoerrgthkhgxvht/app.bsky.feed.post/3kbqnek5ffv2f</guid></item></channel></rss>