Towards AI @towards_AI

Making AI accessible to all with our courses, blogs, tutorials, books & community. Since 2019 we helped teach over 400k AI. Now we also help corporate teams. academy.towardsai.net United States Joined April 2019

Tweets

23K
Followers

41K
Following

2K
Likes

2K

Towards AI @towards_AI

8 minutes ago

Not much sign of an end of the LLM scaling paradigm with OpenAI expecting to spend $19bn on training next year!

Louie Peters @_LouiePeters

8 minutes ago

Not much sign of an end of the LLM scaling paradigm with OpenAI expecting to spend $19bn on training next year!

0 1 4 369 3

Download Image

0 0 3 16 0

Louie Peters @_LouiePeters

a day ago

One of the biggest mistakes in LLM Development/AI engineering - particularly for enterprise projects - is not bringing the industry expertise of your company / end users deeply into the development process. Building successful LLM workflows/agents/chatbots should really be a…

swyx 🇸🇬 @swyx

a day ago

98 56 1K 280K 781

Download Image

0 1 4 1K 3

Louie Peters @_LouiePeters

a day ago

I find so few personal AI use cases for ChatGPT/Gemini/Grok where i would prefer to trade off intelligence (or maximum context length) for lower cost / latency. It feels so natural now to just set a task running and switch tabs to work on other things. So much more efficient than…

0 1 9 134K 6

Louis-François Bouchard 🎥🤖 @Whats_AI

a day ago

I quit my PhD in Artificial Intelligence. At the time, it felt like a risky decision. But what’s changed since then, 1.5 years later? Honestly—a lot. 🧵👇

2 1 2 330 1

Download Image

Towards AI @towards_AI

2 days ago

🚀 GPT-5 is now #1 in coding On Aider’s Polyglot benchmark (225 tasks across C++, Go, Java, JS, Python, Rust), GPT-5 scores 88.0%, beating o3-pro (84.9%) and Gemini 2.5 Pro (83.1%). Note: this benchmark run cost about $29 vs $146 for o3-pro at high reasoning GPT-5 is not only…

0 0 1 189 2

Download Image

Alex Albert @alexalbert__

3 days ago

A conversation with Claude Code's creator @bcherny on the future of agentic coding, the evolution of coding models, and how he uses Claude Code to build Claude Code.

48 90 788 69K 614

Download Video

Towards AI @towards_AI

4 days ago

How to actually succeed with enterprise AI ✅ Forget “open vs closed.” The future is hybrid. Frontier APIs on secure enterprise terms for complex tasks + local models for the most sensitive data or latency sensitive workflows = scalable, secure, and smart.

1 0 2 155 1

Download Video

Louis-François Bouchard 🎥🤖 @Whats_AI

4 days ago

I’m thrilled to finally share that we ( @towards_AI ) are launching the TAI Partner Program 🚀 Our goal with this big update is to allow anyone to earn by referring the course. The idea is simple: help one person learn AI, and earn something concrete from the start! 1 referral…

0 1 2 440 0

Download Image

Towards AI @towards_AI

5 days ago

The real cost of running giant open-weight LLMs 💸 Serving a trillion-parameter mixture of experts LLM in-house isn’t just hard — it’s expensive. Secure LLM APIs often win on both cost and capability. Design systems that use both open and closed AI.

0 0 0 139 1

Download Video

Louie Peters @_LouiePeters

5 days ago

My September cohort group call will be on the 7th September, but you can get started in your own time. For a 10% discount valid until 2nd September use code: 24hrdiscount10

0 1 5 542 3

Louie Peters @_LouiePeters

5 days ago

Learn to master using AI for professional work tasks and join the group that benefits from AI! Develop ChatGPT, Gemini or Copilot skills well beyond prompting and find AI workflows that fit your daily routine. Our 15 AI engineers have built the most advanced AI course out there:

1 137 369 396K 67

Towards AI @towards_AI

6 days ago

The enterprise AI trap 🚨 It’s often a mistake to assume you can’t use closed LLM at all for security reasons The best enterprise systems use a combination of secure closed LLM deployments and smaller open models.

1 0 0 177 0

Download Video

Towards AI @towards_AI

6 days ago

GPT-5 just outperformed doctors on the U.S. Medical Licensing Exam (USMLE) 🩺🤯 According to a new research paper, GPT-5 showed massive leaps in medical reasoning: 🧠 On MedXpertQA (multimodal), GPT-5 beat pre-licensed doctors by +24.2% in reasoning and +29.4% in understanding.…

0 0 5 332 0

Download Image

Towards AI @towards_AI

7 days ago

China is dominating the open-weight AI race 🇨🇳⚡” DeepSeek V3.1 is just the latest proof: Chinese labs now own the open-weight leaderboard, while US labs stick to small, safe releases to compliment their closed model business.