Harsha Bandi retweeted
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then:
- the human iterates on the prompt (.md)
- the AI agent iterates on the training code (.py)
The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc.
github.com/karpathy/autorese…
Part code, part sci-fi, and a pinch of psychosis :)
Harsha Bandi retweeted
Replying to @antirez
Great post. Design and discovery are expensive. Reimplementation is now cheap. Will we find ourselves in a world that disincentivizes novel creation?
Harsha Bandi retweeted
lovable build competition with $40k in prizes is starting tomorrow
apply to the competition + repost this post and we'll reach out to you with 100 lovable credits to build with
Harsha Bandi retweeted
Some people today are discouraging others from learning programming on the grounds AI will automate it. This advice will be seen as some of the worst career advice ever given. I disagree with the Turing Award and Nobel prize winner who wrote, “It is far more likely that the programming occupation will become extinct [...] than that it will become all-powerful. More and more, computers will program themselves.” Statements discouraging people from learning to code are harmful!
In the 1960s, when programming moved from punchcards (where a programmer had to laboriously make holes in physical cards to write code character by character) to keyboards with terminals, programming became easier. And that made it a better time than before to begin programming. Yet it was in this era that Nobel laureate Herb Simon wrote the words quoted in the first paragraph. Today’s arguments not to learn to code continue to echo his comment.
As coding becomes easier, more people should code, not fewer!
Over the past few decades, as programming has moved from assembly language to higher-level languages like C, from desktop to cloud, from raw text editors to IDEs to AI assisted coding where sometimes one barely even looks at the generated code (which some coders recently started to call vibe coding), it is getting easier with each step.
I wrote previously that I see tech-savvy people coordinating AI tools to move toward being 10x professionals — individuals who have 10 times the impact of the average person in their field. I am increasingly convinced that the best way for many people to accomplish this is not to be just consumers of AI applications, but to learn enough coding to use AI-assisted coding tools effectively.
One question I’m asked most often is what someone should do who is worried about job displacement by AI. My answer is: Learn about AI and take control of it, because one of the most important skills in the future will be the ability to tell a computer exactly what you want, so it can do that for you. Coding (or getting AI to code for you) is a great way to do that.
When I was working on the course Generative AI for Everyone and needed to generate AI artwork for the background images, I worked with a collaborator who had studied art history and knew the language of art. He prompted Midjourney with terminology based on the historical style, palette, artist inspiration and so on — using the language of art — to get the result he wanted. I didn’t know this language, and my paltry attempts at prompting could not deliver as effective a result.
Similarly, scientists, analysts, marketers, recruiters, and people of a wide range of professions who understand the language of software through their knowledge of coding can tell an LLM or an AI-enabled IDE what they want much more precisely, and get much better results. As these tools are continuing to make coding easier, this is the best time yet to learn to code, to learn the language of software, and learn to make computers do exactly what you want them to do.
[Original text: deeplearning.ai/the-batch/is… ]
Harsha Bandi retweeted
I've just published the 67th edition of my Kafka Monthly Digest. Check it out for updates on 3.6.0, new KIPs, recent community project releases and blogs.
developers.redhat.com/blog/2…
Harsha Bandi retweeted
Backwards Compatibility, Testing Changes to the Go Language and GODEBUG
Yesterday a blog post by @_rsc was released on go.dev. The post is well worth a read but here is a summary of my key takeaways:
- The #golang team takes backwards compatibility really seriously, and will continue to even as Go 2.0 gets created. A go 2.0 program should never break a Go 1.0 program.
- Backwards compatibility is incredibly nuanced. The best example of this is that they wanted to improve the precision of time.Now() but it broke some tests that look like this at Google. They therefore had to take a different approach.
- Whenever a new Go release is created, they test it against the entire Google codebase and use that as an indicator of whether there is any breaking changes. They have also built custom tooling that ensures they never remove exported APIs.
- The Kubernetes team and the Go team work closely together to ensure that any nuances to changes are considered.
- When a new version of Go is ready, Google use it as their production version before the public release.
- If any of the changes the Go team has made does break your code, GODEBUG (go.dev/doc/godebug#history) was formalised in Go 1.21 which means you can essentially "opt-out" of the behaviour changes.
My Take
The Go team being inside Google gives such a large benefit to the #golang community and this post really helps you to understand that. Having the k8s team work so closely with them and having the benefit of one of the largest Go code bases in existence to run tests against is why we get such successful releases regularly.
Thanks for the insight @_rsc
Harsha Bandi retweeted
Top 6 most commonly used Server Types.
1. Web Server:
Hosts websites and delivers web content to clients over the internet
2. Mail Server:
Handles the sending, receiving, and routing of emails across networks
3. DNS Server:
Translates domain names (like bytebytego .com) into IP addresses, enabling users to access websites by their human-readable names.
4. Proxy Server:
An intermediary server that acts as a gateway between clients and other servers, providing additional security, performance optimization, and anonymity.
5. FTP Server:
Facilitates the transfer of files between clients and servers over a network
6. Origin Server:
Hosts central source of content that is cached and distributed to edge servers for faster delivery to end users.
Over to you: Which type of server do you find most crucial in your online experience?
–
Subscribe to our weekly newsletter to get a Free System Design PDF (158 pages): bit.ly/42Ex9oZ
Harsha Bandi retweeted
I'm so excited to announce Laravel Herd - a completely free macOS app to get started with PHP on your system.
No brew required, built-in PHP updates and much more.
We've been working on this for the last couple of weeks and I'm super proud of it.
herd.laravel.com
Harsha Bandi retweeted
We at @Google created FREE comprehensive training on Generative AI! Get ready to apply AI in your workplace.
👉 Intro to Generative AI
👉 Intro to Large Language Models
👉 Intro to Responsible AI
👉 Intro to Image Generation
👉 Encoder-Decoder
👉 Attention Mechanism
👉 Transformer Models and BERT Model
👉 Create Image Captioning Models
👉 Intro to Generative AI Studio
Course Link: cloudskillsboost.google/path…
Blog: cloud.google.com/blog/topics…
#GenAI #MachineLearning #ArtificialIntelligence
Had a terrible experience with @SpiritAirlines at DFW. Staff repeatedly made us remove items from our bags for size, even though we measured them multiple times. Then, when boarding, they checked again. Flight was mostly empty. Staff just played with us.#poorservice #Disappointed
1
1
96
we haven’t faced this issue with our leg of our flight.
No wonder @SpiritAirlines were listed last in customer satisfaction surveys @theACSI @JDPower
1
1
76
Harsha Bandi retweeted
📣 Angular has a new change detection and performance guide, which covers:
‣ Basics of change detection
‣ Most common performance pitfalls
‣ How to identify them with Angular DevTools
‣ Solutions that will help you speed your application up ⬆️
📚 angular.io/guide/change-dete…
Harsha Bandi retweeted
Want find-in-page to work with collapsed content in HTML? 🔎 https://xitter.cf/t.co/mLbPmeTSPS The new hidden=until-found attribute & beforematch auto-opens sections on find. In Chromium 102 thx to @JosephArhar!
Harsha Bandi retweeted
1/ How can I optimize my frontend for the best availability?
I want to guarantee customers will almost never see a broken page.
What would the architecture and infrastructure look like?
Harsha Bandi retweeted
Very clever trick! I’d find this easier to understand, though:
const hasDuplicates = (array) => new Set(array).size < array.length;
xitter.cf/KittyGiraudel/status/1…
Harsha Bandi retweeted
I've made @ChromeDevTools touch-friendly by upgrading mouse events to pointer events (w3c.github.io/pointerevents/). 🧑💻
Now you can resize split panes and adjust colors on touch-screen devices by finger or pen, not only mouse. 👆✍️
#DX #DeveloperExperience #UX #UserExperience
Harsha Bandi retweeted
Why do we think Partytown is such a big deal?
Because third party scripts are fighting your code for main thread resources, and your performance and user experience suffers
Partyown solves this by offloading 3rd party scripts to a web worker in just a couple lines of code
Harsha Bandi retweeted
Happy new year!
Today, a mini-tweetstorm of some of the things I've said and written over the past decade, and what I think about those subjects today.
Harsha Bandi retweeted
I read a lot of books this year to broaden my horizons in AI/ML with adjacent or complementary disciplines. It was a great pleasure so I’m sharing some my reading list here with a couple of notes:
[1/12]
Harsha Bandi retweeted
Merry Christmas! We got you a new telescope.
The James Webb Space Telescope launched today, beginning a one-million-mile journey to see 13.5 billion years into the past. Follow @NASAWebb and join the quest to #UnfoldTheUniverse: go.nasa.gov/3qHIEKT





















![Code snippet: const hasDuplicate = array => array.some((entry, index, array) => array.indexOf(entry) !== index ) hasDuplicate(['k', 'i', 't', 't', 'y']) // true hasDuplicate(['g', 'i', 'r', 'a', 'u', 'd', 'e', 'l']) // false](https://pbs.twimg.com/media/FJoYcKkWQAgamRr.jpg?name=small&format=webp)






