llm

https://aistudio.google.com/
#Google AI Studio是我最强大的AI工具，使用的 #LLM 当然是 #Gemini ，可以在1.0Pro、1.5Pro和1.5flash中选择（Gemini网页和app上的1.5付费用户专享，这里免费不限量），1.5支持1M tokens（2M版本排队中），温度、安全性等参数可调，完全免费，唯一的问题是没有移动页面。
AI Studio是为开发设计的，比面向用户的聊天UI功能丰富得多，当然也复杂一些。
作为用户，这个开发UI有什么用呢？
首先，你能整本书扔给他，甚至一次几本书也行。梳理情节、分析人物、搜寻段落，效果很不错，就像一个伴读。幻觉仍有，人物经历的前半段通常没问题，结局经常出错，输出太长还是容易胡言乱语。这是处理长文本最好的工具，图片、音频、视频也都支持。
还有个结构化提示功能，可以设置多组输入-输出对作为示例，每组中的输入、输出都可以有多个，输入也可以是图片。Gemini会学习这些示例来处理类似问题。

我目前能体验到的还仅仅是文本模式，就已经体会到了 #GPT-4o 的强大。
请依次计算表面积为1的球体、正方体、正四面体的体积，结果用小数表示。
https://chatgpt.com/share/4a746518-63fd-4f39-80b4-ce27820c810b
#4o 的答案简洁干脆，毫无瑕疵。
这个题目就是套公式，毫无难度，但是4o之前没有一个能算对的，除了 #Wolfram Alpha，而它不算 #LLM 。
我一直相信 #数学 能力是LLMs最重要的能力之一，因为数学的本质是 #逻辑 。 #知识 是否丰富其实并不重要，#RAG 可以很好地补充。
这就像大家都想成为知识渊博的聪明人，但是如果只能二选一的话，无知的聪明人总比什么都知道的傻子要好太多，因为知识是容易学习的，而逻辑就难得多。

陪孩子看 #三国 ，看到 #赤壁 之战前， #诸葛亮 背《 #铜雀台赋 》激周瑜抗曹。“揽 #二乔 于东南兮，乐朝夕之与共”这句显然有问题。 #曹植 一代文豪，写赋公开称颂其父欲夺人之妻，实在匪夷所思。这句本为“连二桥于东西兮，若长空之蝃𬟽”，《三国志》裴松之注所录连“二桥”都没有。
这一点一眼就能看出是小说家言，另一点却需要一点考证。《铜雀台赋》作于铜雀台落成之时，是在赤壁之战两年以后。
赤壁之战前， #曹操 真的说过要抢二乔吗？一个比较完善有理的回答大概是这样：
简单结论：这是民间传言，并无实据。然后解释为什么会有这个传言：一是《三国演义》这段的影响，二是杜牧诗“东风不与周郎便，铜雀春深锁二乔”，三是曹操形象不好，还确实喜欢人妻。
这个逻辑并不复杂，#LLM s能理解吗？

2022年11月 #ChatGPT 发布，迅速火出圈。IT界几乎每家公司都号称自己也有这个技术，但大概半年以后，模仿者才逐渐上线，并且至今还在紧紧追赶。
有几个问题：
1、为什么ChatGPT之前没人说，之后大家都说有？
2、如果本来就有，为什么过了半年才拿出产品？
我想之前确实很多公司都有 #LLM 技术，但都是作为一个 #NLP 工具，没人相信能达到这个效果，所以没有投入很多资源。一旦知道这条路能走通，就开始争先恐后地买卡、训练了。
但他们还是低估了从技术到产品这个过程的复杂程度，所以花了很长时间才能面世，而且也一直追不上ChatGPT 。

Postgres as a search engine
https://anyblockers.com/posts/postgres-as-a-search-engine
https://news.ycombinator.com/item?id=41343814

Build a retrieval system with semantic, full-text, & fuzzy search in Postgres to be used as a backbone in RAG pipelines.

We’ll combine 3 techniques:

* full-text search with tsvector
* semantic search with pgvector
* fuzzy matching with pg_trgm

* bonus: BM25

https://en.wikipedia.org/wiki/Okapi_BM25
https://blog.paradedb.com/pages/elasticsearch_vs_postgres
https://news.ycombinator.com/item?id=41173288

#RDBMS #postgres #PostgreSQL #databases #search #SearchEngine #LLM #RAG #BM25

When food becomes inedible to disgusting: On Facebook, AI bots and fake accounts are flooding culinary groups to make a quick buck. An analysis by @franceinfo - https://www.francetvinfo.fr/internet/reseaux-sociaux/facebook/enquete-franceinfo-images-generees-par-ia-pubs-en-masse-arnaques-les-mauvaises-recettes-des-pages-facebook-de-cuisine_6881498.html - pubs en masse, arnaques… Les mauvaises recettes de Facebook.

#AI #LLM #generativeAI #spam #scam #facebook #Meta #bots #fake #cuisine #kochen #culinary #foodBlogger #food #ShittyFoodporn #arnaque

TECH BROs v. THE OCEAN:

CRYPTO: let's boil the oceans to create fake money for criminals. We'll fleece the rubes and make miillions. LOL you own that JPG now. Sure you do.

LLMs: let's boil the oceans to create pure garbage out of people's intellectual property. We'll steal from everyone and make millions.

THE OCEANS: brb making some hurricanes

#crypto #llm #ai #techbros #climate #idiots

DeepSeek launched a free, open-source large-language model in late December, claiming it was developed in just two months at a cost of under $6 million — a much smaller expense than the one called for by Western counterparts.

These developments have stoked concerns about the amount of money big tech companies have been investing in AI models and data centers, and raised alarm that the U.S. is not leading the sector as much as previously believed.

The "Western counterparts" are claiming training a model might take years and billions of dollars. This has always been a hyped-up grift, with snake oil salesmen and con artists being showered with money and power. It's really quite amazing how profoundly unintelligent "the market" is in practice.

The sad reality is that the US could lead in this field (1), if we'd stop routinely putting narcissists and con artists in charge and showering them with praise even when they fail.

From https://www.cnbc.com/2025/01/27/nvidia-falls-10percent-in-premarket-trading-as-chinas-deepseek-triggers-global-tech-sell-off.html

#AI #GenAI #GenerativeAI #LLM #SnakeOil #hype #grift #MarketCapitalism

(1) Putting aside whether we should, which is an important question.

Sabot in the Age of AI

Here is a curated list of strategies, offensive methods, and tactics for (algorithmic) sabotage, disruption, and deliberate poisoning.

🔻 iocaine
The deadliest AI poison—iocaine generates garbage rather than slowing crawlers.
🔗 https://git.madhouse-project.org/algernon/iocaine

🔻 Nepenthes
A tarpit designed to catch web crawlers, especially those scraping for LLMs. It devours anything that gets too close. @aaron
🔗 https://zadzmo.org/code/nepenthes/

🔻 Quixotic
Feeds fake content to bots and robots.txt-ignoring #LLM scrapers. @marcusb
🔗 https://marcusb.org/hacks/quixotic.html

🔻 Poison the WeLLMs
A reverse-proxy that serves diassociated-press style reimaginings of your upstream pages, poisoning any LLMs that scrape your content. @mike
🔗 https://codeberg.org/MikeCoats/poison-the-wellms

🔻 Django-llm-poison
A django app that poisons content when served to #AI bots. @Fingel
🔗 https://github.com/Fingel/django-llm-poison

🔻 KonterfAI
A model poisoner that generates nonsense content to degenerate LLMs.
🔗 https://codeberg.org/konterfai/konterfai

OpenAI 称有证据 DeepSeek 使用「模型蒸馏」技术利用其模型进行训练。

- The Verge 的文章还提到，考虑到是 OpenAI 开盗用互联网数据训练其模型之先河，这一指控颇具有讽刺性。
- 404 Media 的讽刺则更为直接："Hahahahahahahahahahahahahahahaha hahahhahahahahahahahahahahaha"。

theverge.com/~
404media.co/~
seealso: HackerNews:42865527

#LLM #OpenAI #DeepSeek #today

Telegram 原文

If you turn the sink off when you're done using it to conserve water, or turn the lights off when you leave the room to conserve electricity, why do you use ChatGPT or other AI tools? Using those sorts of tools a few times a month negates whatever you've conserved by being prudent in other parts of your life.

#AI #GenAI #GenerativeAI #LLM #waste #environment

Resistance to the coup is the defense of the human against the digital and the democratic against the oligarchic.

From https://snyder.substack.com/p/of-course-its-a-coup

Defense of the human against the digital has been my mission for some time. Resisting the narratives about how #LLMs "reason", "pass the Turing test", "diagnose illnesses", are "better than humans" in various ways are part of it. Resisting the false narrative that we're on the verge of discovering #AGI is part of it. Allowing these false stories to persist and spread means succumbing to very dark anti-human forces. We're seeing some of the consequences now, and we're seeing how far this might go.

#USPol #AI #GenAI #GenerativeAI #LLM #AGI