ทวีตที่ปักหมุด
Ada Boese
547 posts

Ada Boese
@AdaBoese
Learning about LLMs as a I build the best article generator the world has ever seen! 🟩🟨⬜️⬜️⬜️⬜️⬜️⬜️⬜️⬜️ $50k/m
New York เข้าร่วม Ekim 2023
283 กำลังติดตาม149 ผู้ติดตาม
Ada Boese รีทวีตแล้ว

Must bookmark.
@PunkpEye has put together a huge collection of 300+ MCP server implementations tailored for AI agents! 🤯
This Github repo links AI models to real-world tools such as web browsers, APIs, databases, cloud services, CLIs, & much more!
↳ github.com/punkpeye/aweso…

English

@AlexBelogubov It is a tough space/tough problem. I've been working on AIMD app for several months now. Focused on quality beyond anything else. I have good results, but the application got extremely complex to maintain.
English

I am launching a new product - Marevo ✍️
This is an AI Writing Tool with ready-made templates for creating texts.
Write any text content in seconds.
It's live on Product Hunt now 👇
producthunt.com/posts/marevo

English

So far this year, I’ve had $4,464.92 of interest deposited into my savings account.
That’s almost $900/month for free.
I had this money sitting with Chase earning 0.01% for awhile until @carriganisms told me to switch to Wealthfront and upped that to 5.5% 🤑

English

I don't know what changed, but AI generated content is taking off the ground big time this month. I am increasingly using AIMD.app for my own projects rather than focusing on selling it... and holy guacamoly did I see a shift in Google Search Console this month compared to previous.
Google algo changes?
English

@pbteja1998 So much love for this project! Been checking the progress since the start. Amazing how much you've achieved!
English

I launched SiteGPT.ai exactly 1 year ago…
And we just crossed $250k in lifetime revenue today 😀
Hopefully we reach the next $250k in much less than a year.
I have so many ideas on how to take SiteGPT to the next stage…
…and we are just getting started.

English

@lilybankai Thanks for letting me know! You entered Https in the URL instead of https, so it didn’t work :) I fixed it, now it should work in different case of letters, try it now.
English

I'm launching my new product Stealth.Marketing 🥷
Creating individual stories for subreddits with fast native promotion of your project.
This is the promotion method that I actively use for my projects and I decided to automate it.
👉 producthunt.com/posts/stealth-…
English

Super proud to support International Women's Day for the whole weekend😊
Microlaunch.net has the privilege of being a second home for many talented women who make great products👌
Please show your support by giving feedback on their product weekend and help them reach the Top-10🎖
🎬@hannaz_z | building BoolVideo (slug: boolvideo)
💡@Juliet__de | building Fipner AI (slug: flipnerai)
👥@shreya_gr | building AICamp (slug: aicamp)
✨@cyb3ramelia | building Kleanse (slug: kleanseapp)
💸@BusMark_w_Nika | building on YT (slug: businessmarketingwithnika)
🎶@verry_codes | building Drumbitious (slug: drumbitious)
💬@Carollmir | building MojiPulse (slug: mojipulse)
📋@Catalina_Melnic | building Morrow (slug: morrow)
✍🏻@adaboese | building AIMD (slug: aimd)
🦾@EnumHQ | build Enum Smart AI Assistant (slug: enumsmartaiassistant)
How to vote: microlaunch.net/p/
Example: microlaunch.net/p/flipnerai
We definitely need more women in tech and Women's Day should be every day🙏
Have a nice weekend✨

English

Claude vs. OpenAI GPT-4 generated content, side-by-side comparison
- OpenAI GPT-4 gist.github.com/adaboese/12e3c…
- Claude 3 (Opus) gist.github.com/adaboese/d0b73…
Both of these are outputs of AIMD app. They are not made using a single prompt, but rather using RAG with over a dozen instructions. This allows to test a quite broad range of expectations, such as the adherence to instructions, error rate, speed, etc. Since the two model APIs are mostly compatible, I've decided to compare it side-by-side.
A few interesting observations:
- Claude followed instructions a lot closer than OpenAI. The outline that was provided to the initial instructions is pretty close to the final article structure despite multiple revisions.
- Claude output scored better in terms of use of broader set of data formats (tables, lists, quotes).
- Contrary to many tweets, Claude output is not excessively verbose. Worth mentioning that part of RAG instructions to rewrite content for brevity.
- Claude took 5 minutes to execute 52 prompts. OpenAI took 7 minutes.
– Image inputs (prompts) are generated with respective models, but the actual images are generated using DALLE.
English

@mattshumer_ @levelsio This works equally well with OpenAI just in case. That's the approach that I use.
English

@levelsio @AnthropicAI I've not seen anyone talk about throughput speeds. How are those? thinking of trying this out for AIMD.app, although pretty happy with current LLMs
English

Trying Claude 3 today by @AnthropicAI for cbt.chat
At first glance it feels slightly better than GPT4 and way better than Mistral etc.
The first thing I notice is that its responses seem MUCH more human than LLMs before it
"I know it may not feel like it right now, but you ARE going to get through this."
That ARE is quite human, and GPT4 would never capitalize emphasis like that unless you ask it to

@levelsio@levelsio
Added first RAG to cbt.chat it can now open URLs "retrieval augmented generation (RAG)"
English

🔥 We're live on @ProductHunt
Open Graph & Twitter Image Templates for Next.js ▲
Would love your support 🙌
producthunt.com/posts/og-image…
English







