Ken Kuo

[Patently Apple @PatentlyApple]

https://t.co/nSeeZ8MrOb

[Patently Apple @PatentlyApple]

https://t.co/5ZcbuKECca

[Similarweb @Similarweb]

GenAI Traffic Share, Key Takeaways:

→ Grok and DeepSeek continue to regain ground.
→ Claude surpasses Perplexity.
→ ChatGPT continues to lose share.

🗓️ 12 Months Ago:
ChatGPT: 86.6%
Gemini: 5.6%
Perplexity: 2.2%
Claude: 1.9%
Copilot: 1.6%

🗓️ 6 Months Ago:
ChatGPT: 80.1%
DeepSeek: 6.1%
Gemini: 5.9%
Grok: 2.4%
Perplexity: 1.6%
Claude: 1.2%
Copilot: 1.2%

🗓️ 3 Months Ago:
ChatGPT: 79.0%
Gemini: 8.6%
DeepSeek: 3.9%
Grok: 2.3%
Perplexity: 1.8%
Claude: 1.7%
Copilot: 1.2%

🗓️ 1 Month Ago:
ChatGPT: 74.1%
Gemini: 12.9%
DeepSeek: 3.7%
Perplexity: 2.4%
Grok: 2.0%
Claude: 2.0%
Copilot: 1.2%

🗓️ Today:
ChatGPT: 72.3%
Gemini: 13.7%
DeepSeek: 4.2%
Grok: 2.5%
Claude: 2.4%
Perplexity: 2.3%
Copilot: 1.2%

>>

[SemiAnalysis @SemiAnalysis_]

China’s State Council on October 9 approved Order No. 61 of 2025, announcing export controls on certain overseas rare-earth items. This marks the fourth round of rare-earth export restriction efforts; the previous round was on April 8.
(1/8)🧵

[TechEpiphany @TechEpiphanyYT]

AMD vs Intel - Steam CPU Install Share

Almost as if dominating the DIY market for two straight years actually had an effect on the install base.

img by computerbase (hence the wrong label)

[Christopher Mims 🤌 @mims]

The AI infrastructure build-out is so gigantic that in the past 6 months, it contributed more to the growth of the U.S. economy than /all of consumer spending/

The 'magnificent 7' spent more than $100 billion on data centers and the like in the past three months *alone*

1/🧵

[Similarweb @Similarweb]

GenAI Traffic Share Update — main takeaways:

→ Thanks to the release of Grok 4, Grok is back on a positive trajectory.
→ Google continues to take small bites out of OpenAI's share.

🗓️ 12 months ago:
ChatGPT: 84.7%
Google: 8.4%
Claude: 2.1%
Perplexity: 1.5%
Microsoft: 0.9%

🗓️ 6 months ago:
ChatGPT: 86.6%
Google: 5.3%
Perplexity: 1.9%
DeepSeek: 1.6%
Claude: 1.5%
Microsoft: 1.3%

🗓️ 3 months ago:
ChatGPT: 79.8%
DeepSeek: 6.5%
Google: 5.6%
Grok: 2.6%
Perplexity: 1.5%
Claude: 1.3%
Microsoft: 1.2%

🗓️ 1 month ago:
ChatGPT: 78.4%
Google: 8.6%
DeepSeek: 5.0%
Grok: 2.0%
Perplexity: 1.7%
Claude: 1.5%
Microsoft: 1.1%

🗓️ Today:
ChatGPT: 78.3%
Google: 8.7%
DeepSeek: 4.5%
Grok: 2.6%
Perplexity: 1.6%
Claude: 1.5%
Microsoft: 1.2%

>>

[SemiAnalysis @SemiAnalysis_]

China's share of DUV tool shipments from ASML vs rest of the world

[Evan @StockMKTNewz]

Nvidia's $NVDA Revenue by Region

[Similarweb @Similarweb]

Global AI Tracker Highlights

Some significant shifts across various AI sectors:

1. DeepSeek's share in the GenAI category:

0.7% – 3 months ago
9.6% – 1 month ago
7.9% – Today

>>

[WOLF @WOLF_Financial]

Tesla’s profit pales in comparison to other Mag 7 stocks.

[Quartr @Quartr_App]

$AMZN Q4 2024 📦

Revenue +10%
*Online Stores +8%
*Physical Stores +8%
*Third Party Sellers +9%
*Subscriptions +10%
*Advertising +18%
*AWS +19%
EBIT +61%
*margin 11.3% (7.8)
EPS +86%
FCF LTM +4%

[Quartr @Quartr_App]

$ARM Q3 2025

Revenue +19%
*Royalty +23%
*License +14%
Gross Profit +21%
*marg. 98.1% (96.8)
EBIT +22%
*marg. 45% (43.8)
FCF +39%

(Non-GAAP)

[OpenAI @OpenAI]

We’re testing SearchGPT, a temporary prototype of new AI search features that give you fast and timely answers with clear and relevant sources.

We’re launching with a small group of users for feedback and plan to integrate the experience into ChatGPT. https://t.co/dRRnxXVlGh

[Sam Parker 🇺🇲 @SamParkerSenate]

@elonmusk How come they didn't show the one with the jewish number?

[Visual Capitalist @VisualCap]

Charting the Relationship Between Wealth and Happiness, by Country 😊

From the archive: https://t.co/YHtbBxQWQC

[Lisa Su @LisaSu]

Congratulations @Azure on the launch of the Azure AI infrastructure VM series powered by @AMD Instinct MI300X! We are celebrating this milestone with you and appreciate your partnership, @SatyaNadella! #MSBuild

[Andrew Ng @AndrewYNg]

This week, Google announced a doubling of Gemini Pro 1.5's input context window from 1 million to 2 million tokens, and OpenAI released GPT-4o, which generates tokens 2x faster and 50% cheaper than GPT-4 Turbo and natively accepts and generates multimodal tokens. I view these developments as the latest in an 18-month trend. Given the improvements we've seen, best practices for developers have changed as well.

Since the launch of ChatGPT in Nov 2022, with key milestones that include the releases of GPT-4, Gemini 1.5 Pro, Claude 3 Opus, and Llama 3-70B, many model providers have improved their capabilities in two important ways: (i) reasoning, which allows LLMs to think through complex concepts and and follow complex instructions; and (ii) longer input context windows.

The reasoning capability of GPT-4 and other advanced models makes them quite good at interpreting complex prompts with detailed instructions. Many people are used to dashing off a quick, 1- to 2-sentence query to an LLM. In contrast, when building applications, I see sophisticated teams frequently writing prompts that might be 1 to 2 pages long (my teams call them “mega-prompts”) that provide complex instructions to specify in detail how we’d like an LLM to perform a task. I still see teams not going far enough in terms of writing detailed instructions. For an example of a moderately lengthy prompt, take a look at Claude 3’s system prompt. It’s detailed and gives clear guidance on how Claude should behave.

This is a very different style of prompting than we typically use with LLMs’ web user interfaces, where we might dash off a quick query and, if the response is unsatisfactory, clarify what we want through repeated conversational turns with the chatbot.

Further, the increasing length of input context windows has added another technique to the developer’s toolkit. GPT-3 kicked off a lot of research on few-shot in-context learning. For example, if you’re using an LLM for text classification, you might give a handful — say 1 to 5 examples — of text snippets and their class labels, so that it can use those examples to generalize to additional texts. However, with longer input context windows — GPT-4o accepts 128,000 input tokens, Claude 3 Opus 200,000 tokens, and Gemini 1.5 Pro 1 million tokens (2 million just announced in a limited preview) — LLMs aren’t limited to a handful of examples. With many-shot learning, developers can give dozens, even hundreds of examples in the prompt, and this works better than few-shot learning.

When building complex workflows, I see developers getting good results with this process:
- Write quick, simple prompts and see how it does.
- Based on where the output falls short, flesh out the prompt iteratively. This often leads to a longer, more detailed, prompt, perhaps even a mega-prompt.
- If that’s still insufficient, consider few-shot or many-shot learning (if applicable) or, less frequently, fine-tuning.
- If that still doesn’t yield the results you need, break down the task into subtasks and apply an agentic workflow.

I hope a process like this will help you build applications more easily. If you’re interested in taking a deeper dive into prompting strategies, I recommend Microsoft's Medprompt paper (Nori et al., 2023), which lays out a complex set of prompting strategies that can lead to very good results.

[Original text (with links): https://t.co/UOtLDza1Vh ]

[OpenAI @OpenAI]

We're opening up access to our new flagship model, GPT-4o, and features like browse, data analysis, and memory to everyone for free (with limits). https://t.co/H2dB6w7Ynz

[OpenAI @OpenAI]

Live demo of GPT-4o realtime conversational speech