ChatGPT search, from OpenAI

 Introducing ChatGPT search | OpenAI

"ChatGPT can now search the web in a much better way than before. You can get fast, timely answers with links to relevant web sources, which you would have previously needed to go to a search engine for. This blends the benefits of a natural language interface with the value of up-to-date sports scores, news, stock quotes, and more.

ChatGPT will choose to search the web based on what you ask, or you can manually choose to search by clicking the web search icon."

OpenAI’s search engine is now live in ChatGPT - The Verge

"ChatGPT is officially an AI-powered web search engine. The company is enabling real-time information in conversations for paid subscribers today (along with SearchGPT waitlist users), with free, enterprise, and education users gaining access in the coming weeks.

Rather than launching as a separate product, web search will be integrated into ChatGPT’s existing interface. The feature determines when to tap into web results based on queries, though users can also manually trigger web searches. ChatGPT’s web search integration finally closes a key competitive gap with rivals like Microsoft Copilot and Google Gemini, which have long offered real-time internet access in their AI conversations."

tiny web server in C

shenfeng/tiny-web-server: a tiny web server in C, for daily use. @GitHub

it works! from 2013; 440 lines of code

just seem to keep CPU very busy...

it is educational, but there is a reason for production-grade web servers...

MIT license

DraganSr: The simplest local web file server !

"production" grade web server in Go is 14 lines of code!

not surprising, since Go language is co-developed by same person who co-developed C language, 40 years before!

Ken Thompson - Wikipedia

In 1972, Ritchie completed the original version of the high-level C programming language; Thompson then rewrote the UNIX kernel in C. This made UNIX portable — that is, easily adapted for different computer platforms and allowed it to be used almost without change across a wide span of computers.

Inductee Ken Thompson and the History of C Programming Language

The Future, as predicted by Elon Musk

Elon Musk Goes on a Rampage of Stunning Predictions (Full) - YouTube

Elon Musk interview with Peter Diamandis

Cloud Native: Everything Fails All the Time

 ‘Everything Fails All the Time’ – Communications of the ACM

Amazon vice president and chief technology officer Werner Vogels has kept the company at the forefront of scalable systems since he joined in 2004. Vogels was one of the main driving forces behind the architecture of Amazon Web Services (AWS). His insistent reminders that "everything fails all the time" have been hugely influential to the developers of high-availability systems. Here, he talks about continuing to innovate in database technology—and keeping up with customer needs.

Now Go Build with Werner Vogels – S1E7 Porto | Amazon Web Services - YouTube

Chris Klug — Active Solution Play — Active Solution

search with Postgres & pgvector

 Hybrid search with PostgreSQL and pgvector | Jonathan Katz

in the context of vector search, “hybrid search” is the act of combining an alternative searching method with a vector similarity search. Hybrid search uses multiple search methods over the same data, performs a ranking of the results for each search method, and then combines all the results to determine a final ranking before returning the results. Hybrid search is often used to improve the quality of the returned results, or in other words, boost the “recall” rate.

This blog post will introduce what hybrid search is and how to use it with PostgreSQL and pgvector.

EV: Tesla 48V: Low-Voltage Connector Standard

 Tesla Drops Another BOMBSHELL Standard! - YouTube

Tesla launches low-voltage connector standard to simplify EV transition

Tesla has officially launched a new standard for the low-voltage connections in electric vehicles (EVs), which the company says will reduce necessary connection types in most EVs from over 200 to just six.

In addition, Tesla writes that the LVCS suite was designed upon the 48-volt architecture built into the Cybertruck, meeting certain requirements for spacing for 48V operation. The company notes that the 48V architecture requires just a quarter of the current to deliver the same amount of power as commonly-used 12V systems.

...substantially reduced parts and wiring needed in the truck’s low-voltage wiring network.

Standardizing Automotive Connectivity | Tesla

Midjourney: bootstrap AI, $200M

Apparently, good and profitable AI models and companies could me made without billions of dollars, like in case of OpenAI

Midjourney.com

Midjourney - Wikipedia

Midjourney is a generative artificial intelligence program and service created and hosted by the San Francisco-based independent research lab Midjourney, Inc. Midjourney generates images from natural language descriptions, called prompts, similar to OpenAI's DALL-E and Stability AI's Stable Diffusion.

The Weird Rise Of Anti-Startups - YouTube

Vipul Vaibhaw on X: "Midjourney 's exceptional case study
- - Completely bootstrapped to ~$200 million revenue.
 - First researcher had no experience, she was an independent researcher. Now works at Stability AI.
- David Holz, founder was also CTO/founder of leap motion. He was very much versed"
/ X

Midjourney’s $200M without external investors | LinkedIn

Pieter Levels: Programming, Viral AI Startups, and Digital Nomad Life | Lex Fridman Podcast #440 - YouTube

Pieter Levels (aka levelsio on X) is a self-taught developer and entrepreneur who has designed, programmed, launched over 40 startups, many of which are highly successful.

external HDD vs SDD

  Western Digital My Passport (6TB) review: spacious and stocky | Tom's Hardware 

After years of capacity stagnation at 5TB, Western Digital recently upped its top-capacity portable hard drives to 6TB

the thickness of WD's My Passport drives increases at every capacity pump — starting at 0.44 inches for 1TB, and nearly doubling that (0.81 inches) for the 6TB model

Overall, the 6TB WD drive feels a bit chunky and hefty (it weighs nearly half a pound) for an external hard drive, and it's positively massive compared to most modern external SSDs. Crucial's X9 and X10 drives are about 1/6th the size, though they do top out at 4TB. Samsung's T5 Evo SSD is about half the size and houses up to 8TB of storage. But the Samsung drive is in another price league (at around $600)

you'll pay extra for the highest possible capacity. As of this writing,
the 6TB model sells for $180, while the 5TB model sells for $127.

The capacity jump comes with an increase in z-height, with the 6TB model being 0.81 inches thick to the 5TB's 0.75 inches. So it's likely that the company simply added an extra platter over the 5TB model we looked at back in 2020. 

you can buy a version of this drive with a USB-C cable on the drive side, but it costs $5 more and only comes in gray

Amazon.com: WD 6TB My Passport, Portable External Hard Drive, Black, Backup Software with Defense Against ransomware, and Password Protection, USB 3.1/USB 3.0 Compatible - WDBR9S0060BBK-WESN : Electronics

$185

Amazon.com: WD 6TB Elements Portable External Hard Drive for Windows, USB 3.2 Gen 1/USB 3.0 for PC & Mac, Plug and Play Ready - WDBHJS0060BBK-WESN : Electronics

$174

Amazon.com: WD 5TB My Passport, Portable External Hard Drive, Black, backup software with defense against ransomware, and password protection, USB 3.1/USB 3.0 compatible - WDBPKJ0050BBK-WESN : Electronics

$129  

WD 5TB My Passport, Portable External Hard Drive, Amazon price drop alerts | camelcamelcamel.com

Amazon.com: WD 6TB Elements Portable External Hard Drive for Windows, USB 3.2 Gen 1/USB 3.0 for PC & Mac, Plug and Play Ready - WDBHJS0060BBK-WESN : Electronics

$125

Amazon.com: Western Digital 6TB My Book Desktop External Hard Drive, USB 3.0, External HDD with Password Protection and Auto Backup Software - WDBBGB0060HBK-NESN : Everything Else

$289

WD 16TB My Book Desktop USB 3.0 External Hard WDBBGB0160HBK-NESN  @B&H

$289

AI cryptocurrency? Worldcoin Orb; World ID

 Worldcoin - Wikipedia

World Network, originally Worldcoin, is a for-profit cryptocurrency project that uses iris biometrics developed by San Francisco- and Berlin-based Tools for Humanity. Founded in 2019 by OpenAI chief executive Sam Altman, Max Novendstern, and Alex Blania, it has accumulated $250 million in funding from venture capital firms Andreessen Horowitz and Khosla Ventures, as well as from Reid Hoffman.

Worldcoin's business is to provide a reliable way to authenticate humans online, which it calls World ID. 

Worldcoin hopes to provide a reliable way to authenticate humans online called World ID,^[18] to counter bots and fake virtual identities facilitated by artificial intelligence. 

Orb-shaped iris scanners

Sam Altmans New AI Device Unleashes AI Agents (New Worldcoin Orb) - YouTube

Oracle cloud: nuclear power for AI?

 One of the world's largest software firms is building a data center with not one but three nuclear reactors — Oracle wants to build world's most powerful supercomputer ever, yes, even more powerful than Colossus | TechRadar

"In a bid to meet the spiraling energy demands of advanced AI models, Oracle is set to build a data center powered by three small modular nuclear reactors.

Today, Oracle has 162 cloud data centers, live and under construction throughout the world. The largest of these data centers is 800 megawatts, and it will contain acres of Nvidia GPU clusters able to train the world's largest AI models," Ellison noted.

...the cost of that training gets to be astronomical. When I talk about building gigawatt or multi-gigawatt data centers… the entry price for a real frontier model from someone who wants to compete in that area is around $100 billion."

Micorosft GenAIScript

"demoware" or a real tool?

Generative AI Scripting | GenAIScript

GenAIScript is a scripting language that integrates LLMs into the scripting process using a simplified JavaScript syntax. It allows users to create, debug, and automate LLM-based scripts.

Getting Started | GenAIScript

AI Software Should be More Like Plain Old Software | SIGPLAN Blog

song: AGI (Artificial God Incarnate)

AGI (Artificial God Incarnate) - LIVE - YouTube

by Forrest Brazeal


Forrest Brazeal - YouTube

Forrest Brazeal | LinkedIn

The origin of AI (& LISP)

from "AI newsletter" Forward Future AI (2024-10-25)

🧑‍🚀 AI Whistleblower Alleges OpenAI Violated Copyright Laws

Who coined the term “artificial intelligence?”

The term "artificial intelligence" was coined by computer scientist John McCarthy in 1956. He introduced the term during the Dartmouth Conference, a summer research project at Dartmouth College, which is considered the founding event of AI as a field of study. McCarthy is often referred to as one of the "fathers of AI" for his contributions to the field.

John McCarthy (computer scientist) - Wikipedia

John McCarthy (September 4, 1927 – October 24, 2011) was an American computer scientist and cognitive scientist. He was one of the founders of the discipline of artificial intelligence.^[1] He co-authored the document that coined the term "artificial intelligence" (AI), developed the programming language family Lisp, significantly influenced the design of the language ALGOL, popularized time-sharing, and invented garbage collection.

In the late 1950s, McCarthy discovered that primitive recursive functions could be extended to compute with symbolic expressions, producing the Lisp programming language.^[16] That functional programming seminal paper also introduced the lambda notation borrowed from the syntax of lambda calculus in which later dialects like Scheme based its semantics. Lisp soon became the programming language of choice for AI applications after its publication in 1960.

Lisp (programming language) - Wikipedia

John McCarthy began developing Lisp in 1958 while he was at the Massachusetts Institute of Technology (MIT). McCarthy published its design in a paper in Communications of the ACM in April 1960,

McCarthy's original notation used bracketed "M-expressions" that would be translated into S-expressions. As an example, the M-expression car[cons[A,B]] is equivalent to the S-expression (car (cons A B))

Lisp was first implemented by Steve Russell on an IBM 704 computer using punched cards.

AI: Microsoft Copilot Agent Builder

AI News: Copilot Agent Builder, IBM Granite 3.0, NEW Claude Sonnet, Open-Source Text To Video! - YouTube

Microsoft Copilot Studio | Customize or Create Copilots

"Enhance Microsoft 365 Copilot with agents or build your own custom experiences.

Competition is good :) Agentforce!

Salesforce CEO Marc Benioff Says Microsoft AI Is Clippy 2.0 | Entrepreneur

"Microsoft rebranding Copilot as ‘agents’? That’s panic mode. Let’s be real—Copilot’s a flop because Microsoft lacks the data, metadata, and enterprise security models to create real corporate intelligence. That is why Copilot is inaccurate, spills corporate data, and forces…"

OpenAI: No one is ready for AGI

 No one is ready for AGI — not even OpenAI - The Verge

Miles Brundage, OpenAI’s senior adviser for the readiness of AGI (aka human-level artificial intelligence), delivered a stark warning as he announced his departure on Wednesday: no one is prepared for artificial general intelligence, including OpenAI itself.

“Neither OpenAI nor any other frontier lab is ready [for AGI], and the world is also not ready,”

OpenAI stung by departure of yet another AI safety expert, and fresh claims around copyright infringement | Fortune

Stack Overflow AI: "Knowledge-as-a-service"

New marketing acronym :KaaS, after SaaS :) 

Knowledge-as-a-service: The future of community business models - Stack Overflow

"When enterprises combine this public knowledge store with their knowledge, this newly expanded corpus and its access and maintenance become knowledge-as-a-service,"

classes: Stanford CS101, CS106: Computing, Programming

CS101 Introduction to Computing Principles

CS101 Syllabus

CS101: Introduction to Computing Principles teaches the essential ideas of computers to an audience with no prior computer experience. CS101 is geared to use live code exercises in class -- bringing the ideas to life, but without bogging down too much in computer idiosyncrasies.

If you would like to learn how to code, you should take CS106A. If, instead, you're interested in learning more about how computers work or the technical concepts discussed in the news, CS101 is the class for you!

useful links to "notes" and "reading materials"
using subset of JavaScript for code examples


For high schools or other outside groups, a version of the materials are freely available at introcomputing.org.



CS106A | Summer 24
CS106A: Programming Methodologies (with Python)

Code in Place
Take the first half of CS106A, online for free, with the support from thousands of teachers.

OpenAI API : Functions vs Tools vs Structured Outputs vs Assistants

similar functionality, different name = unnecessary confusion

Functions VS Tools - What is the difference? - API - OpenAI Developer Forum

functions have been deprecated, and tools are the way forward. Tools: A list of tools the model may call. Currently, only functions are supported as a tool.

API Reference - OpenAI API

functions

Deprecated

array

Optional

Deprecated in favor of tools.

API Reference - OpenAI API

tools

array

Optional

A list of tools the model may call. Currently, only functions are supported as a tool. Use this to provide a list of functions the model may generate JSON inputs for. A max of 128 functions are supported.

Assistants overview - OpenAI API

• Assistants can call OpenAI’s models with specific instructions to tune their personality and capabilities.
• Assistants can access multiple tools in parallel. These can be both OpenAI-hosted tools — like code_interpreter and file_search — or tools you build / host (via function calling).
• Assistants can access persistent Threads. Threads simplify AI application development by storing message history and truncating it when the conversation gets too long for the model’s context length. You create a Thread once, and simply append Messages to it as your users reply.
• Assistants can access files in several formats — either as part of their creation or as part of Threads between Assistants and users. When using tools, Assistants can also create files (e.g., images, spreadsheets, etc) and cite files they reference in the Messages they create.

Assistant	Purpose-built AI that uses OpenAI’s models and calls tools
Thread	A conversation session between an Assistant and a user. Threads store Messages and automatically handle truncation to fit content into a model’s context.
Message	A message created by an Assistant or a user. Messages can include text, images, and other files. Messages stored as a list on the Thread.
Run	An invocation of an Assistant on a Thread. The Assistant uses its configuration and the Thread’s Messages to perform tasks by calling models and tools. As part of a Run, the Assistant appends Messages to the Thread.
Run Step	A detailed list of steps the Assistant took as part of a Run. An Assistant can call tools or create Messages during its run. Examining Run Steps allows you to introspect how the Assistant is getting to its final results.

Assistant Tools - OpenAI API

openai/openai-assistants-quickstart: OpenAI Assistants API quickstart with Next.js. @GitHub

Function Calling - OpenAI API

As an alternative to function calling you can instead constrain the model's regular output to match a JSON Schema of your choosing. Learn more about when to use function calling vs when to control the model's normal output by using response_format.

When you turn it on by setting strict: true, in your function definition, Structured Outputs ensures that the arguments generated by the model for a function call exactly match the JSON Schema you provided in the function definition.

computers: von Neumann architecture

von Neumann architecture - Wikipedia

a computer architecture based on a 1945 description by John von Neumann, and by others, in the First Draft of a Report on the EDVAC.^[1] The document describes a design architecture for an electronic digital computer with these components:

• A processing unit with both an arithmetic logic unit and processor registers
• A control unit that includes an instruction register and a program counter
• Memory that stores data and instructions
• External mass storage
• Input and output mechanisms

See also: Stored-program computer and Universal Turing machine § Stored-program computer

John von Neumann - Wikipedia

How the Computer was Accidentally Invented - YouTube by Newsthink - YouTube

book

The Man from the Future: The Visionary Life of John von Neumann: Bhattacharya, Ananyo: 9781324003991: Amazon.com: Books

AI Raspberry Pi 5, with GPU and AI Kit (NPU)

 Every Developer Needs a Raspberry Pi - YouTube

Raspberry Pi 8gb, 4gb, and 2gb models: https://www.seeedstudio.com/Raspberry... $80
Raspberry Pi 5 PCIe to M.2 NVMe 2CH: https://s.click.aliexpress.com/e/_Dd4...
l298n motor controller: https://s.click.aliexpress.com/e/_DlQ...

Amazon.com: Raspberry Pi 5, Single Board Computer, 4GB RAM, 2.4GHz 64-bit Quad-core Arm Cortex-A76 Processor, Bluetooth 5.0, BLE Wireless : Electronics

$90 

Buy a Raspberry Pi AI Kit – Raspberry Pi

The AI module is a 13 tera-operations per second (TOPS) neural network inference accelerator built around the Hailo-8L chip.

Raspberry Pi AI Kit for Pi 5 @CanaKit $70

ts-blank-space: TypeScript to JavaScript

ts-blank-space

ts-blank-space is a fast type-stripping compiler that converts TypeScript to JavaScript. It supports a modern subset of TypeScript by erasing the types and replacing them with whitespace. That's it. It is not a type checker and does not perform any other code transformations.

The underlying technique can improve build performance and simplify development. The implementation is pure TypeScript. It is simple enough to read and understand in a few minutes because it is only 700 lines of code and reuses the original TypeScript parser.

bloomberg/ts-blank-space: A small, fast, pure JavaScript type-stripper that uses the official TypeScript parser. @GitHub

AI: "Machines of Loving Grace" by Anthropic CEO

interesting blog post from founder/CEO of one most prominent AI companies,
(in size of an essay or small book)

Dario Amodei — Machines of Loving Grace

"most people are underestimating just how radical the upside of AI could be &
most people are underestimating how bad the risks could be."

define "loving grace" - Google Search

"Loving grace is the idea that love is unconditional,
and that it is given to people when they least deserve it"

so much about not "anthropomorphising" 
(attribute human characteristics or behavior to, a god, animal, or object).


also introducing term "powerful AI" instead "AGI"

• In terms of pure intelligence⁴, it is smarter than a Nobel Prize winner across most relevant fields – biology, programming, math, engineering, writing, etc. This means it can prove unsolved mathematical theorems, write extremely good novels, write difficult codebases from scratch, etc.

• Basic assumptions and framework
• 1. Biology and health
• 2. Neuroscience and mind
• 3. Economic development and poverty
• 4. Peace and governance
• 5. Work and meaning
• Taking stock

Dario Amodei - Wikipedia

Dario Amodei (born 1983) is an Italian-American artificial intelligence researcher and entrepreneur. He is the co-founder and CEO of Anthropic, the company behind the large language model series Claude AI. He was previously the vice president of research at OpenAI.

Caltech, Stanford University, Princeton University (PhD)

Anthropic - Wikipedia

Anthropic PBC is a U.S.-based artificial intelligence (AI) public-benefit startup founded in 2021. It researches and develops AI to "study their safety properties at the technological frontier" and use this research to deploy safe, reliable models for the public.^[5][6][7] Anthropic has developed a family of large language models (LLMs) named Claude as a competitor to OpenAI's ChatGPT and Google's Gemini.

QuickJS Javascript Engine

QuickJS Javascript Engine (C, MIT license, updated 2024)

QuickJS is a small and embeddable Javascript engine. It supports the ES2023 specification including modules, asynchronous generators, proxies and BigInt.

It optionally supports mathematical extensions such as big decimal floating point numbers (BigDecimal), big binary floating point numbers (BigFloat) and operator overloading.

Small and easily embeddable: just a few C files, no external dependency, 210 KiB of x86 code for a simple hello world program.

Fast interpreter with very low startup time: runs the 76000 tests of the ECMAScript Test Suite in less than 2 minutes on a single core of a desktop PC. The complete life cycle of a runtime instance completes in less than 300 microseconds.

Almost complete ES2023 support including modules, asynchronous generators and full Annex B support (legacy web compatibility). Passes nearly 100% of the ECMAScript Test Suite tests when selecting the ES2023 features

bellard/quickjs: Public repository of the QuickJS Javascript Engine. @GitHub

Fabrice Bellard's Home Page

Fabrice Bellard - Wikipedia

Fabrice Bellard (French pronunciation: [fa.bʁis bɛ.laʁ]; born 1972) is a French computer programmer known for writing FFmpeg, QEMU, and the Tiny C Compiler. He developed Bellard's formula for calculating single digits of pi.

Tiny CC, from JavaScript, with Bun

TCC : Tiny C Compiler (LGPL)

• SMALL! You can compile and execute C code everywhere, for example on rescue disks (about 100KB for x86 TCC executable, including C preprocessor, C compiler, assembler and linker).
• FAST! tcc generates x86 code. No byte code overhead. Compile, assemble and link 9 times faster than GCC.
• UNLIMITED! Any C dynamic library can be used directly. TCC is heading torward full ISOC99 compliance. TCC can of course compile itself.
• SAFE! tcc includes an optional memory and bound checker. Bound checked code can be mixed freely with standard code.
• Compile and execute C source directly. No linking or assembly necessary. Full C preprocessor and GNU-like assembler included.

Compile and run C in JavaScript | Bun Blog

From compression to cryptography to networking to the web browser you're reading this on, the world runs on C. If it's not written in C, it speaks the C ABI (C++, Rust, Zig, etc) and is available as a C library. C and the C ABI are the past, present, and future of systems programming.


from: JavaScript Weekly Issue 705: September 19, 2024

Now You Can Compile and Run C in JavaScript — Well, if you’re using Bun, at least. Bun v1.1.28 lands experimental support for compiling and native C code and then running its functions from JavaScript. It's easier than it sounds..

TinyCC/tinycc: Unofficial mirror of mob development branch @GitHub

author: Fabrice Bellard - Wikipedia

Fabrice Bellard (French pronunciation: [fa.bʁis bɛ.laʁ]; born 1972) is a French computer programmer known for writing FFmpeg, QEMU, and the Tiny C Compiler. He developed Bellard's formula for calculating single digits of pi.

Index of /releases/tinycc/ downloads, including Windows build

OTCC : Obfuscated Tiny C Compiler

the smallest C compiler which is able to compile itself. 

The original OTCC code could only run on i386 Linux

MicroPython

 micropython/micropython: MicroPython - a lean and efficient Python implementation for microcontrollers and constrained systems @GitHub

C, MIT license

This is the MicroPython project, which aims to put an implementation of Python 3.x on microcontrollers and small embedded systems. You can find the official website at micropython.org.

MicroPython implements the entire Python 3.4 syntax (including exceptions, with, yield from, etc., and additionally async/await keywords from Python 3.5 and some select features from later versions). 

MicroPython - Python for microcontrollers

Overview — MicroPython latest documentation

PythonImplementations - Python Wiki