Analytics Camp

𝗪𝗵𝘆 𝗠𝗼𝗿𝗲 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 𝗠𝗶𝗴𝗵𝘁 𝗡𝗼𝘁 𝗕𝗲 𝗕𝗲𝘁𝘁𝗲𝗿
There is a common belief that if one AI agent is good, then ten must be better. I’ve been looking into some recent research from Google that suggest this isn't always true. In some cases, adding more agents can actually make the overall system less effective. It is a good reminder that we can't just throw more computing power at a problem and expect it to fix itself. We need to be smarter about how we design these workflows rather than just making them more complex. I broke down those shocking results in this video: https://youtu.be/tzloGPymubI

Here's the paper: Towards a science of scaling agent systems: arxiv.org/abs/2512.08296

#AIAgents #GoogleAI #TechTrends #SoftwareEngineering #AIResearch

Skip the Carousel

1 2 3

1 day ago | [YT] | 3

Analytics Camp

Hope every one of you are well in this super crazy world situation! I'm publishing the new video today; this is about Google's very interesting findings in their recent paper. If you've been struggling to make your AI agents projects work properly, you will definitely need to know this: https://youtu.be/tzloGPymubI

Why your AI agents is failing and how to fix it - based on Google's important research paper: Towards a Science of Scaling Agent Systems. All important links in the video's decsription box. THANK YOU FOR YOUR ATTENTION TO THIS MATTER

#aiagents #agenticai #google #multi-agentsystems #mas

Skip the Carousel

1 2 3

5 days ago | [YT] | 2

Analytics Camp

Hey everyone! I hope every one of you are doing well in this super cold season. In 2026 I'm going to work on various generative AI models, research, & new tech and on the side I intend to continue with different aspects of AI voice generation tutorials for some time. So I thought I give you a brief overview of AI voice tutorials I've done in 2025 in one place. These are beginner-friendly tutorials so give them a go:

1) Expressive AI voice design with emotional tags using Maya1 plus model quantisation for low VRAM on modest hardware:
Watch Me Build The Perfect AI Voice w/ Maya1 | Emotional Voice Design Tutorial with quantisation for low VRAM 👉
https://youtu.be/naDDr_ABU4M

2) Realistic AI voice /audio engineering using 5 natural language methods with Kokoro-TTS free models:
THIS Trick Makes AI Voices Sound Human: FREE podcast-style kokoro TTS Tutorial 👉 https://youtu.be/nwELsTaELSM

3) Fundamentals of Text to Speech (TTS) and voice cloning, tutorial + demo:
I cloned Andrew Ng’s Voice in 10 Lines of Code: Super-realistic FREE Text-To-Speech (TTS)
https://youtu.be/XcAL34MEdMA

4) An important summary of main advancements, models, and methods in AI voice / audio generation in 2025 and the models and methods that you should expect in 2026 + a summary of best tips and tricks for Voice AI:
👉 https://youtu.be/UavKSnCgjk0

1 month ago | [YT] | 1

Analytics Camp

I summarised the most important advancements in AI voice generation ( text to speech / TTS, voice cloning, voice design and translation) in 2025 in this video, along with what we should expect in AI voice / audio generation in 2026.
https://youtu.be/UavKSnCgjk0?si=WesH3...

I also discussed what I learned from building a ton of AI voice workflows and what I would do in these workflows in 2026 if I'd started over again. Let me know which model was your favourite one this year.

3 months ago | [YT] | 2

Analytics Camp

Watch Me Build The Perfect AI Voice with Maya1 | Emotional Voice Design Tutorial 👉 https://youtu.be/naDDr_ABU4M

After a whole month of wrestling with my hardware/software issues, today I released the coolest voice design engineering with the Maya1 AI voice model; 100% free and open source, easy loading with 4bit quantisation optimised for low VRAM, full SNAC voice decoder explained, and a demo of a multi-turn, multi-speaker podcast-style AI conversation designed entirely with Maya1.

Don't forget to leave your comments below the video to help me design my next videos better. Thanks and happy watching :)

#AI #aivoicegeneration #texttospeech #maya1 #freeaivoice

3 months ago | [YT] | 4

Analytics Camp

Good news: Voice AI series is continuing! I'm working on a more comprehensive video about voice design with audio features & voice engineering. The aim is to include as many desired emotions and voice features as we need 'directly' inside the script with special tags and the model will engineer the EXACT voice we want, such as 'excitement', 'scream', 'curiosity', or even 'anger', besides many more emotional expressions.

What's more, we will see how we can 'describe' a speaker's voice with features such as 'podcast-style', 'professional', 'friendly voice', and 'low-pace', as well as defining accents (e.g., American, British, Middle Eastern, etc) and many more voice features to give our AI a realistic human-like voice!

If this is interesting, stay tuned, the video's coming soon :)

Meanwhile, if you haven't watched the previous two Voice AI videos, here's a quick playlist for both: www.youtube.com/playlist?list...

#AI #VoiceAI #VoiceGeneration #TextToSpeech

Skip the Carousel

1 2

4 months ago | [YT] | 4

Analytics Camp

Around a month ago Kimi Team introduced Kimi Linear (& KDA Attention Mechanism). It seems it has the potential to replace traditional LLMs (Attention mechanism in Transformers). Some also believe it could make an LLM paradigm-shift. Honestly, its a bit heavy; the model is about 98GB; so I can make a video about it if this post gets enough likes (I don't know, maybe more than 20 likes!). Or just comment and let me know if you want a paper/architecture walk-through and/or a code tutorial.

Here's a simplified explanation of the paper in case you're interested: www.linkedin.com/feed/update/urn:li:activity:73944…

#LLMs #AI #transformers #attention #KimiLinear

4 months ago | [YT] | 6

Analytics Camp

I have included the two code tutorials about Text to Speech (TTS) and voice cloning in a playlist for your convenience 👉 www.youtube.com/playlist?list...

I'm planning to publish more TTS and voice cloning videos using the latest and best free models. If you're interested in a similar topic, please let me know. Have a great rest of your day :)

#ai #tts #texttospeech #voiceai #voicecloning

Skip the Carousel

1 2

4 months ago | [YT] | 9

Analytics Camp

Just in case you haven't noticed, I released an important tutorial video on Text-To-Speech (TTS) with free voice models from Kokoro. I show how to use certain voice/audio features to generate natural and realistic AI voices for even very long texts & multi-speaker, multi-lingual workflows. Here's the link, please leave your comments: https://youtu.be/nwELsTaELSM

#tts #ai #texttospeech #voiceai #freeaivoice

Skip the Carousel

1 2

4 months ago | [YT] | 2

Analytics Camp

𝐇𝐨𝐰 𝐃𝐢𝐝 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐌𝐨𝐝𝐞𝐥𝐬 𝐂𝐨𝐦𝐞 𝐓𝐨 𝐓𝐡𝐢𝐬 𝐏𝐨𝐢𝐧𝐭?
Trying to explain the history behind language models to non-technical viewers, I realised how mystifying language models seem to most people! This inspired me to create an explanation that absolutely anyone can understand - no background needed. The analogies will click even for your grandparents! If you've ever wondered what's actually happening inside these AI systems, this is for you 👉 https://youtu.be/n_5spvz-2KI

#LanguageModels #AIexplained #Beginners #LLMs #AI

Skip the Carousel

1 2 3

5 months ago | [YT] | 2