Milan Vasarhelyi
@vasarmilan.bsky.social
Writing about automations, cutting-edge AI models and tools.
I automate workflows of e-commerce, agency and real estate companies.
Automations Starter Pack: https://go.bsky.app/L9SRGJ
I automate workflows of e-commerce, agency and real estate companies.
Automations Starter Pack: https://go.bsky.app/L9SRGJ
Pinned
Milan Vasarhelyi
@vasarmilan.bsky.social
· Nov 21
Starter pack for people working, in or talking about, workflow automation:
DM or comment if you want to be on it, or know someone who should be on it!
go.bsky.app/L9SRGJK
DM or comment if you want to be on it, or know someone who should be on it!
go.bsky.app/L9SRGJK
Today I realized that FancyZones, the built-in tiling window manager of Windows can actually do 90% of what i3 can.
Honestly Windows, at least for dev, came a very long way when I last used it, and I don't regret switching back from Linux a while ago at all.
Honestly Windows, at least for dev, came a very long way when I last used it, and I don't regret switching back from Linux a while ago at all.
January 2, 2025 at 4:08 PM
Today I realized that FancyZones, the built-in tiling window manager of Windows can actually do 90% of what i3 can.
Honestly Windows, at least for dev, came a very long way when I last used it, and I don't regret switching back from Linux a while ago at all.
Honestly Windows, at least for dev, came a very long way when I last used it, and I don't regret switching back from Linux a while ago at all.
A lesson for anyone doing development as a freelancer / agency:
If progress is slower than expected, sometimes saying an "I didn't forget you" email to a client is more important than grinding an hour more.
If progress is slower than expected, sometimes saying an "I didn't forget you" email to a client is more important than grinding an hour more.
December 13, 2024 at 2:14 PM
A lesson for anyone doing development as a freelancer / agency:
If progress is slower than expected, sometimes saying an "I didn't forget you" email to a client is more important than grinding an hour more.
If progress is slower than expected, sometimes saying an "I didn't forget you" email to a client is more important than grinding an hour more.
LLM coding errors are usually user errors.
Specifically errors with not having enough context.
If you use a recent/obscure library, you should include the API reference in the prompt. And also the relevant parts of your code.
And then either o1 or Claude will give first-time perfect code 90%.
Specifically errors with not having enough context.
If you use a recent/obscure library, you should include the API reference in the prompt. And also the relevant parts of your code.
And then either o1 or Claude will give first-time perfect code 90%.
December 12, 2024 at 1:17 PM
LLM coding errors are usually user errors.
Specifically errors with not having enough context.
If you use a recent/obscure library, you should include the API reference in the prompt. And also the relevant parts of your code.
And then either o1 or Claude will give first-time perfect code 90%.
Specifically errors with not having enough context.
If you use a recent/obscure library, you should include the API reference in the prompt. And also the relevant parts of your code.
And then either o1 or Claude will give first-time perfect code 90%.
When I ask o1 Pro to reformat some text for me
That I could probably do with "find and replace" as well
That I could probably do with "find and replace" as well
rick and morty are sitting at a table talking about their purpose
ALT: rick and morty are sitting at a table talking about their purpose
media.tenor.com
December 11, 2024 at 2:25 PM
When I ask o1 Pro to reformat some text for me
That I could probably do with "find and replace" as well
That I could probably do with "find and replace" as well
The way everyone searches the web will definitely change over the next few years.
For some users it already did.
The question is: will Google be leading this change or will it become irrelevant?
For some users it already did.
The question is: will Google be leading this change or will it become irrelevant?
December 10, 2024 at 9:39 AM
The way everyone searches the web will definitely change over the next few years.
For some users it already did.
The question is: will Google be leading this change or will it become irrelevant?
For some users it already did.
The question is: will Google be leading this change or will it become irrelevant?
Every time I have to copy a password from one screen to another, I mix up a capital i with a smal L.
December 10, 2024 at 9:34 AM
Every time I have to copy a password from one screen to another, I mix up a capital i with a smal L.
Don’t always automate processes immediately.
First, take some time to document and optimize them.
If you automate them, you freeze inefficiency.
And you’re missing a big opportunity to improve.
First, take some time to document and optimize them.
If you automate them, you freeze inefficiency.
And you’re missing a big opportunity to improve.
December 9, 2024 at 1:48 PM
Don’t always automate processes immediately.
First, take some time to document and optimize them.
If you automate them, you freeze inefficiency.
And you’re missing a big opportunity to improve.
First, take some time to document and optimize them.
If you automate them, you freeze inefficiency.
And you’re missing a big opportunity to improve.
I got the $200 ChatGPT Pro.
But I'm thinking about going back to the $20 next month.
I use it 4 hours a day for writing code, I consider it my main tool, so I don't see the price unreasonable.
But the truth is that right now the benefit of o1 pro over o1 is marginal.
But I'm thinking about going back to the $20 next month.
I use it 4 hours a day for writing code, I consider it my main tool, so I don't see the price unreasonable.
But the truth is that right now the benefit of o1 pro over o1 is marginal.
December 9, 2024 at 1:47 PM
I got the $200 ChatGPT Pro.
But I'm thinking about going back to the $20 next month.
I use it 4 hours a day for writing code, I consider it my main tool, so I don't see the price unreasonable.
But the truth is that right now the benefit of o1 pro over o1 is marginal.
But I'm thinking about going back to the $20 next month.
I use it 4 hours a day for writing code, I consider it my main tool, so I don't see the price unreasonable.
But the truth is that right now the benefit of o1 pro over o1 is marginal.
You can access the best AIs for $20/mo plans
But that changes very soon.
As OpenAI launched ChatGPT Pro, for $200/month.
With exclusive access to o1-pro.
Will you be getting it?
But that changes very soon.
As OpenAI launched ChatGPT Pro, for $200/month.
With exclusive access to o1-pro.
Will you be getting it?
December 6, 2024 at 10:01 AM
You can access the best AIs for $20/mo plans
But that changes very soon.
As OpenAI launched ChatGPT Pro, for $200/month.
With exclusive access to o1-pro.
Will you be getting it?
But that changes very soon.
As OpenAI launched ChatGPT Pro, for $200/month.
With exclusive access to o1-pro.
Will you be getting it?
Results of ChatGPT on Advent of Code (daily programming quizzes) in 2023 vs 2024, so far.
AI shines in small and isolated problems while humans are still much better at working in a complex system, which is the bulk of SW engineering.
But for tasks like this, AI had insane progress since 2023.
AI shines in small and isolated problems while humans are still much better at working in a complex system, which is the bulk of SW engineering.
But for tasks like this, AI had insane progress since 2023.
December 6, 2024 at 9:17 AM
Results of ChatGPT on Advent of Code (daily programming quizzes) in 2023 vs 2024, so far.
AI shines in small and isolated problems while humans are still much better at working in a complex system, which is the bulk of SW engineering.
But for tasks like this, AI had insane progress since 2023.
AI shines in small and isolated problems while humans are still much better at working in a complex system, which is the bulk of SW engineering.
But for tasks like this, AI had insane progress since 2023.
Sometimes a one-word change can lead to a 94% cost saving in an AI project.
For a classification app where we prototyped with GPT-4o, with 75% accuracy...
Switching to 4o-mini still kept at 74%.
Cost went down from $2.50 to $0.150 / 1M tokens.
Optimizing the prompt was 60%➡️75% accuracy.
#ai
For a classification app where we prototyped with GPT-4o, with 75% accuracy...
Switching to 4o-mini still kept at 74%.
Cost went down from $2.50 to $0.150 / 1M tokens.
Optimizing the prompt was 60%➡️75% accuracy.
#ai
December 5, 2024 at 9:28 AM
Sometimes a one-word change can lead to a 94% cost saving in an AI project.
For a classification app where we prototyped with GPT-4o, with 75% accuracy...
Switching to 4o-mini still kept at 74%.
Cost went down from $2.50 to $0.150 / 1M tokens.
Optimizing the prompt was 60%➡️75% accuracy.
#ai
For a classification app where we prototyped with GPT-4o, with 75% accuracy...
Switching to 4o-mini still kept at 74%.
Cost went down from $2.50 to $0.150 / 1M tokens.
Optimizing the prompt was 60%➡️75% accuracy.
#ai
Classify & Summarize 60 000 websites for just $50 in API costs?
We implemented a cost-effective workflow to mass-label websites for a #SEO agency offering #backlink services.
This way, they'll always have a list of websites that they can reach out to in any niche & geography.
#Airtable & #Openai
We implemented a cost-effective workflow to mass-label websites for a #SEO agency offering #backlink services.
This way, they'll always have a list of websites that they can reach out to in any niche & geography.
#Airtable & #Openai
December 4, 2024 at 4:50 PM
Reposted by Milan Vasarhelyi
What are the best starter packs you’ve found?
December 4, 2024 at 2:53 AM
What are the best starter packs you’ve found?
I never had success in using AI for copywriting, but I think I now found a hack.
I used it to help write a landing page. But I prompted it to have a conversation, not one go.
Giving some suggestions for just one section, one paragraph each, refining with feedback, and only then going to the next.
I used it to help write a landing page. But I prompted it to have a conversation, not one go.
Giving some suggestions for just one section, one paragraph each, refining with feedback, and only then going to the next.
December 3, 2024 at 2:50 PM
I never had success in using AI for copywriting, but I think I now found a hack.
I used it to help write a landing page. But I prompted it to have a conversation, not one go.
Giving some suggestions for just one section, one paragraph each, refining with feedback, and only then going to the next.
I used it to help write a landing page. But I prompted it to have a conversation, not one go.
Giving some suggestions for just one section, one paragraph each, refining with feedback, and only then going to the next.
It's too easy to get lost in scrolling a social media feed while working...
So I made a Chrome extension to make it just a little harder.
Now looking at each post is a decision and not automatic.
So I made a Chrome extension to make it just a little harder.
Now looking at each post is a decision and not automatic.
December 2, 2024 at 1:13 PM
It's too easy to get lost in scrolling a social media feed while working...
So I made a Chrome extension to make it just a little harder.
Now looking at each post is a decision and not automatic.
So I made a Chrome extension to make it just a little harder.
Now looking at each post is a decision and not automatic.
As useful as no-code automation can be
I love that in Airtable I could just write a JS script to do some data processing on the cells, or call some APIs.
Usually it's faster than setting up a Make/Zapier automation.
#nocode #airtable #automation
I love that in Airtable I could just write a JS script to do some data processing on the cells, or call some APIs.
Usually it's faster than setting up a Make/Zapier automation.
#nocode #airtable #automation
November 26, 2024 at 12:30 PM
As useful as no-code automation can be
I love that in Airtable I could just write a JS script to do some data processing on the cells, or call some APIs.
Usually it's faster than setting up a Make/Zapier automation.
#nocode #airtable #automation
I love that in Airtable I could just write a JS script to do some data processing on the cells, or call some APIs.
Usually it's faster than setting up a Make/Zapier automation.
#nocode #airtable #automation
Reposted by Milan Vasarhelyi
Still plenty of room in my 👶🏻 NEWCOMERS starter pack.
Less than 1000 followers? Indiehacker building something? Please like and comment below 👇
You don't have to share, but it helps us all if you do ❤️
go.bsky.app/75HfkVp
Less than 1000 followers? Indiehacker building something? Please like and comment below 👇
You don't have to share, but it helps us all if you do ❤️
go.bsky.app/75HfkVp
November 24, 2024 at 10:31 PM
Still plenty of room in my 👶🏻 NEWCOMERS starter pack.
Less than 1000 followers? Indiehacker building something? Please like and comment below 👇
You don't have to share, but it helps us all if you do ❤️
go.bsky.app/75HfkVp
Less than 1000 followers? Indiehacker building something? Please like and comment below 👇
You don't have to share, but it helps us all if you do ❤️
go.bsky.app/75HfkVp
Hosting a #webinar but drowning in busywork? 4 ways to automate most of it:
1) Auto-register attendees via Eventbrite + Zoom.
2) Automate reminders + follow-ups.
3) Log registrations in Sheets for analysis
4) Sync attendees to your CRM and add them to nurturing flows.
1) Auto-register attendees via Eventbrite + Zoom.
2) Automate reminders + follow-ups.
3) Log registrations in Sheets for analysis
4) Sync attendees to your CRM and add them to nurturing flows.
November 22, 2024 at 9:48 AM
Hosting a #webinar but drowning in busywork? 4 ways to automate most of it:
1) Auto-register attendees via Eventbrite + Zoom.
2) Automate reminders + follow-ups.
3) Log registrations in Sheets for analysis
4) Sync attendees to your CRM and add them to nurturing flows.
1) Auto-register attendees via Eventbrite + Zoom.
2) Automate reminders + follow-ups.
3) Log registrations in Sheets for analysis
4) Sync attendees to your CRM and add them to nurturing flows.
Starter pack for people working, in or talking about, workflow automation:
DM or comment if you want to be on it, or know someone who should be on it!
go.bsky.app/L9SRGJK
DM or comment if you want to be on it, or know someone who should be on it!
go.bsky.app/L9SRGJK
November 21, 2024 at 4:32 PM
Starter pack for people working, in or talking about, workflow automation:
DM or comment if you want to be on it, or know someone who should be on it!
go.bsky.app/L9SRGJK
DM or comment if you want to be on it, or know someone who should be on it!
go.bsky.app/L9SRGJK
If you're interested in e-commerce, check out this starter pack by @gui.do : go.bsky.app/C4DLifZ
November 20, 2024 at 12:18 PM
If you're interested in e-commerce, check out this starter pack by @gui.do : go.bsky.app/C4DLifZ
Reposted by Milan Vasarhelyi
SuperAnnotate wants to help companies manage their AI data sets
SuperAnnotate wants to help companies manage their AI data sets
SuperAnnotate, a platform for managing and prepping AI training data sets, has raised a new round of capital.
© 2024 TechCrunch. All rights reserved. For personal use only.
tcrn.ch
November 18, 2024 at 2:03 PM
SuperAnnotate wants to help companies manage their AI data sets
TikTok's parent company, ByteDance, released a new portrait generation model.
They can match the nuances of any human emotion and expression better than similar models before.
This is expected to be a step up for animation and movie production use-cases.
They can match the nuances of any human emotion and expression better than similar models before.
This is expected to be a step up for animation and movie production use-cases.
November 19, 2024 at 11:22 AM
TikTok's parent company, ByteDance, released a new portrait generation model.
They can match the nuances of any human emotion and expression better than similar models before.
This is expected to be a step up for animation and movie production use-cases.
They can match the nuances of any human emotion and expression better than similar models before.
This is expected to be a step up for animation and movie production use-cases.
Anthropic's Claude will now be available to US intelligence and defense agencies.
Anthropic is the most safety-minded out of all major LLM makers, and the last one to make this move.
The models will run on the Palantir’s AI on AWS.
Anthropic is the most safety-minded out of all major LLM makers, and the last one to make this move.
The models will run on the Palantir’s AI on AWS.
November 19, 2024 at 11:16 AM
Anthropic's Claude will now be available to US intelligence and defense agencies.
Anthropic is the most safety-minded out of all major LLM makers, and the last one to make this move.
The models will run on the Palantir’s AI on AWS.
Anthropic is the most safety-minded out of all major LLM makers, and the last one to make this move.
The models will run on the Palantir’s AI on AWS.
In a court case, a long thread of emails were released in which the idea of OpenAI formed...
Very useful read to understand what went on behind the curtains.
Elon wanted OpenAI to be strongly attached to Tesla while keeping fully non-profit.
The others did not.
#ElonMusk #OpenAI #SamAltman #AI
Very useful read to understand what went on behind the curtains.
Elon wanted OpenAI to be strongly attached to Tesla while keeping fully non-profit.
The others did not.
#ElonMusk #OpenAI #SamAltman #AI
November 19, 2024 at 11:11 AM
In a court case, a long thread of emails were released in which the idea of OpenAI formed...
Very useful read to understand what went on behind the curtains.
Elon wanted OpenAI to be strongly attached to Tesla while keeping fully non-profit.
The others did not.
#ElonMusk #OpenAI #SamAltman #AI
Very useful read to understand what went on behind the curtains.
Elon wanted OpenAI to be strongly attached to Tesla while keeping fully non-profit.
The others did not.
#ElonMusk #OpenAI #SamAltman #AI