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.
You might like
🚀 Land your dream LLM job in just 40 hours! AI & LLM engineers are in really high demand, but the skills gap holds many back. If you have some Python experience, we’ll teach you PRECISELY what employers want through hands-on projects and portfolio-building. All you need is…
Very rare glimpse inside LLM architecture at closed AI labs. Grok-3 and Grok-4 were 3 trillion total parameters, Grok-5 will be 6 trillion. This still doesn't tell us active parameters for the models (given Mixture of Experts architectures) but it sounds like Grok-5 will be…
Huge Leaks on Grok-5 and its predecessors from recent Elon Musk interview: - "Grok-5 is a 6 trillion parameter model, whereas Grok 3 and 4 are based on a 3 trillion parameter model" - "the 6 trillion parameters will have a much higher intelligence density per gigabyte than…
We are looking for AI engineers! Ideally, we'd need a French- and English-speaking candidate, but we also hire English-only candidates. If you have built cool products around LLMs and are interested in AI education, giving training AND building actual industry solutions, please…
AI updates are cool and all, but nothing hits like the moment it actually stops using em dashes. Pure freedom.
GPT-5 has been upgraded to GPT-5.1. 𝗜𝗻𝘀𝘁𝗮𝗻𝘁 is now warmer, more conversational, and better at following the exact instructions you give. On harder questions it can pause and think more deeply instead of replying fast and guessing. 𝗧𝗵𝗶𝗻𝗸𝗶𝗻𝗴 got smarter about time,…
There are several key things missed in this surface level AI GPU lifespan due diligence and debate. Both AI model architectures and AI chip design are complex optimisation problems. For model design you can trade off cost, latency, vision capability, text capability, reasoning…
Understating depreciation by extending useful life of assets artificially boosts earnings -one of the more common frauds of the modern era. Massively ramping capex through purchase of Nvidia chips/servers on a 2-3 yr product cycle should not result in the extension of useful…
Early Nano Banana 2 findings are in from an insider, these are the favorites after two weeks. Different build than the one that got leaked recently, This one is likely a later checkpoint. The quality is ridiculous, near perfect prompt matching. Follow to find out when it…
It’s frustrating how labs like @Kimi_Moonshot, @Alibaba_Qwen, @deepseek_ai, @allen_ai, @huggingface... share their research, pipelines, and lessons openly, only for closed-source labs to quietly use that knowledge to build better models without ever giving back.
Real-life Real Steel is no longer sci-fi 🤖 Robots are now learning and mimicking human movement live, in real time. The future of robotics just got a lot closer.
@xai Some feedback and product request! @grok s agentic web search via API is a standout product and super useful to have multiple rounds packaged into one API call. The x access is also great for some of the workflow/agentic research pipeline we are building for clients. I also…
Tired of slow, vague, or confidently wrong AI answers? We wrote a guide with @Neo Kim that outlines the 33 must-know LLM concepts. Understand tokens, embeddings, context windows, RLHF, RAG, and guardrails so you can get better results, reduce hallucinations, and make smart…
While this is an x post of a blog post of a threads post; this story still feels feasible to me and just one of many many cases where LLMs are already generating orders of magnitude more value than their cost
A guy just used @AnthropicAI Claude to turn a $195,000 hospital bill into $33,000. Not with a lawyer. Not with a hospital admin insider. With a $20/month Claude Plus subscription. He uploaded the itemized bill. Claude spotted duplicate procedure codes, illegal “double…
You all remember the Taylor Swift scandal. That wasn't a simple filter—it was advanced LLMs creating fakes that were indistinguishable from reality. 🤯 If this tech is already in the hands of the public, just imagine what's happening inside major AI labs. The real danger isn't…
In the AI capex debate it is often missed just how much room for further growth we still have. The big tech AI buildout funders (MSFT, Amazon, Google, Oracle, META, Tesla) have EBITDA of $671bn in 2025 and net debt of just $110bn (nearly all Oracle). At a 19% CAGR the group's…
These two charts tell quite a story on AI data center capex! The increased compute (and capability unlock) due to test time token scaling from reasoning models and agents caught the industry by surprise. These charts do not even yet capture @Sama / @OpenAI recently raising the…
Anthropic’s Claude can now execute tasks, not just describe them. 👀 With Agent Skills, it runs scripts, follows instructions, and performs complex steps — like an actual digital coworker. 💻 Get the full story every week in the Towards AI newsletter. 📩
PewDiePie revealed “ChatOS,” his self built local AI setup at home. The system runs on ten GPUs, shown as a mixed rack of NVIDIA cards, and he uses it to host large language models entirely on his own hardware. In the video he demonstrates Llama 70B and Alibaba’s Qwen models…
24 hours left. Not for a discount. For a seat in a room where people are actually building AI systems. Here’s what’s happening 👇 Tomorrow, we close the November Full Stack AI Engineering cohort. On November 2, we get on a live call and start working through the exact playbook…
AI capex has escalated rapidly to ~$500bn rate and an increasingly large amount is also off balance sheet via firm GPU rental deals and minority JVs. This may only be the start though; OpenAI alone has lined up 30 GW and $1.4trn of spend. Sam targets 250GW by 2033 which would…
“CAPEX IS A PROBLEM!” Says the bears. Everyone has been calling this for years and yet we are at all time highs. Here is the expected spend for 2025. $AMZN $125B $GOOGL $93B $MSFT $85B $META $72B AI Infrastructure and ASIC chip production is costly, but this is a once in a…
United States Trends
- 1. Good Saturday 29.8K posts
- 2. GAME DAY 11.9K posts
- 3. Massie 78.9K posts
- 4. #SaturdayVibes 4,406 posts
- 5. #Caturday 3,801 posts
- 6. #Varanasi 62.7K posts
- 7. #MeAndTheeSeriesEP1 876K posts
- 8. Senior Day 2,317 posts
- 9. Willie Green 1,831 posts
- 10. Draymond 27.3K posts
- 11. #GlobeTrotter 425K posts
- 12. PONDPHUWIN AT MAT PREMIERE 600K posts
- 13. Marjorie 93K posts
- 14. Brooklynn 1,408 posts
- 15. St. Albert N/A
- 16. Diosa 7,630 posts
- 17. Bubba 68.7K posts
- 18. Wemby 50.8K posts
- 19. America Only 34.7K posts
- 20. Alice Wong 1,624 posts
You might like
-
1ONLINEHOST - Professional web hosting 2025
@1onlinehost -
Innovative Binhu
@InnovativeBinhu -
CDT
@China_Datang -
Explore Pingyang
@ExplorePingyang -
CNBM Papua New Guinea
@CNBMPNG -
Verity Systems
@VSSPltd -
Tongxiang China
@Tongxiang_China -
Visit Yinzhou
@VisitYinzhou -
In Zunyi
@InZunyi -
Mazen Turkmani - Gaming News
@MithrieTV -
Crypto Valley Journal (English)
@CVJournal_ch_EN -
Louis-François Bouchard 🎥🤖
@Whats_AI -
Discover Taicang
@DiscoverTaicang -
Sheamus McG
@sheamusmcgov -
Jilin, China
@JilinChina
Something went wrong.
Something went wrong.