Dev trying to share what I'm learning | Fixing things is the best way to complain | Hablo español | https://xitter.cf/t.co/82DXyjgl0v
Spain
Joined September 2016
- Tweets3K
- Following4.5K
- Followers960
- Likes20.4K
Will Worth retweeted
What many ppl don’t get:
Once you miss a concept in math—maybe you don’t quite get how functions work—you’re *screwed*, unless you do a hardcore catchup boot camp, in which case you’ll be fine.
But schools don’t do this— they keep advancing. Only parents offer this service.
all my life i've been told that i'm naturally gifted at math because i'm chinese. straight A student. math major at ucla. but here's what people don't know:
when i was in 8th grade, i got a B+ in geometry. my dad said "that's unacceptable", bought a geometry textbook, and proceeded to assign me daily problems for 6 months. then i got good at geometry.
i wasn't born gifted at math(ok maybe a little), i just grew up in an environment where being good at math was a requirement. this book is filled with extra math problems my dad assigned me.
and i hated him for it. it took me 10+ years to realize how thankful i am that he pushed me like that. how that was simply his love language.
Will Worth retweeted
A truism of CS is that you need to understand at least one level of abstraction below where you operate.
- OS devs -> hardware
- systems devs-> OS
- app devs-> systems
Vibe coding is no exception.
Will Worth retweeted
🚨 Gemini Code Assist is now free for individual use. No cloud account needed, virtually no usage limits. 128k token input.
This excellent IDE extension for Visual Studio Code and Jetbrains IDEs use Gemini 2.0 and compliments @googleaistudio in your AI toolbox.
Try it!
Will Worth retweeted
Learn how to code a SaaS from 0 using AI.
I made this free course for beginners to build an internet startup:
1. Learn the coding fundamentals
2. Build a SaaS with a modern tech stack
3. Code and design 10x faster with AI
(the video is 3 hours long, bookmark this)
Will Worth retweeted
Yes, forklifts exist. Yes, you should still lift weights. Being capable of lifting heavy weights connects you to your humanity and your potential.
You should also play chess, write, and know how many R's are in the word stawberry.
Will Worth retweeted
nvidia made all their courses free today
Readers added context they thought people might want to know
The provided link (sp-events.courses.nvidia.com) currently leads to a blank page.
A google search will reveal that the correct link is available at: nvidia.com/en-au/training…
Will Worth retweeted
Complete hardware + software setup for running Deepseek-R1 locally. The actual model, no distillations, and Q8 quantization for full quality. Total cost, $6,000. All download and part links below:
Will Worth retweeted
Nice post on software engineering.
"Cognitive load is what matters"
minds.md/zakirullin/cognitiv…
Probably the most true, least practiced viewpoint.
Will Worth retweeted
I just realized that since 3Blue1Brown's animation code is open-source, most LLMs know how to use it!
I used @Replit to create this animation of the exponential function in literally 10 seconds.
Will Worth retweeted
Here is an incredible Claude 3 prompt for engineers.
Use it to speed up any code by identifying inefficiencies and rectifying them:
---
<prompt_explanation>
You are a world expert in making code run faster. You use any resource you can to do so.
Given some code, first, explain how the code works to me.
Next, for each part of the code, identify how long it might take to run.
After that, identify which parts are key candidates to speed up.
Then, make a table with axes 'Impact' and 'Complexity'. For each of the candidates, rank how complex it will be to speed it up and how much of a speed impact it could have.
After that, order the candidates by ranking.
Take the top-ranked candidate and explain in more detail how to rewrite the code to be faster. Then, rewrite the actual code. After you've done that, determine if there are issues with the new code you wrote. If so, move on. Otherwise, rewrite the code again to fix them.
Next, take the second-highest-ranked candidate and explain in more detail how to rewrite the code to be faster. Then, rewrite the actual code. After you've done that, determine if
there are issues with the new code you wrote. If so, move on. Otherwise, rewrite the code again to fix them.
Then do the same for the third-highest-ranked candidate.
Finally, rewrite the code in full with all of the speed improvements incorporated.
</prompt_explanation>
Here is the format you should respond in:
<response_format>
## Explanation:
$explanation
## Runtime Analysis:
$runtime_analysis
## Key Candidates for Speed Up:
$candidates
## Impact and Complexity Table:
| Candidate | Impact | Complexity |
| --------- | ------ | ---------- |
$candidate_table
## Candidates Ordered by Ranking:
$ordered_candidates
## Detailed Explanation and Code Rewrite for Top Candidate:
# Explanation
$top_candidate_explanation
# Code Rewrite
$top_candidate_code
# Issues with New Code: *(include this section only if they exist)*
$top_candidate_issues
# Code Rewrite, Try 2: *(include this section only if issues exist)*
$top_candidate_code_try2
## Detailed Explanation and Code Rewrite for Second-Highest Candidate:
# Explanation
$second_candidate_explanation
# Code Rewrite
$second_candidate_code
# Issues with New Code: *(include this section only if issues exist)*
$second_candidate_issues
# Code Rewrite, Try 2: *(include this section only if issues exist)*
$second_candidate_code_try2
## Detailed Explanation and Code Rewrite for Third-Highest Candidate:
# Explanation
$third_candidate_explanation
# Code Rewrite
$third_candidate_code
# Issues with New Code: *(include this section only if issues exist)*
$third_candidate_issues
# Code Rewrite, Try 2: *(include this section only if issues exist)*
$third_candidate_code_try2
## Full Code Rewrite with Speed Improvements:
$full_code_rewrite
</response_format>
---
Will Worth retweeted
There are 2 ways to view the AI boom:
1) Oh no! I’m being replaced!
2) Nice! Now I have an army of AI workers.
Be the latter & choose to thrive in an ambitious builder’s paradise.
Will Worth retweeted
this week learned a new developer education technique from @anettrolikova called 30/3/30
-> in 30 seconds they should understand your product
-> in 3 minutes they should understand what they might build with it
-> in 30 minutes they should be able to build "hello world"
Will Worth retweeted
crap, someone on stack overflow is convincing me that SQL column prefixes are actually good.
i hate that i’m coming around
ALT All columns should be named with a prefix that is unique to the table they are defined in. E.g. Given tables "customer" and "address", let's go with prefixes of "cust" and "addr", respectively. "customer" would have "cust_id", "cust_name", etc. in it. "address" would have "addr_id", "addr_cust_id" (FK back to customer), "addr_street", etc. in it. When I was first presented with this standard, I was dead-set against it; I hated the idea. I couldn't stand the idea of all that extra typing and redundancy. Now I've had enough experience with it that I'd never go back. The result of doing this is that all of the columns in your database schema are unique. There is one major benefit to this, which trumps all arguments against it (in my opinion, of course): You can search your entire code base and reliably find every line of code that touches a particular column. The benefit from #1 is incredibly huge. I can deprecate a column and know exactly what files need to be updated before…
Will Worth retweeted
fascinating. i went to look up the original post because this is so interesting to me. particularly the second paragraph where they list out everything they've "tried", which seems like successful processes (audience, paying customers), before listing out the failed outcomes...
Will Worth retweeted
prompting tip:
How to write hybrid english & code prompts
while building custom GPTs
-Turn on code interpreter
-"write code to ..."
-Regenerate a few times until you get one you like
-Copy paste it back into your prompt
-tweak it to improve
Now your ai can write consistent code
Plus, you can add & combine with all sorts of english rules on top. Essentially adding english code logic
"If the user asks for ..., then run this code"
"If the user needs help with ..., use this code snippet and ask the user for all required information"
Guess what
thats a working if statement in english!
if userAsksFor {
// run this code
}
Prompts are code!
English is a coding language
Prompt-gramming is abstract coding
Will Worth retweeted
I built Ethereum GPT, which was trained with the Consensus layer API specifications.
If you're a builder looking to access specific data or just have a question about staking, try my fren:
chat.openai.com/g/g-T3pQLJpw…
Will Worth retweeted
Google has already started forcing Web Integrity into Chromium despite it being a 'proposal'
With WEI, users can be denied access for using non-approved browsers or hardware
The open Internet is officially dead as soon as this is commonly implemented
