Today we're taking a look at the author channel by Max Votek.
Max is an entrepreneur, investor, and co-founder of Customertimes.
For 17 years, he has been implementing technologies into the business processes of large companies and living alongside alligators in sunny Florida.
Max offers honest reflections on AI, business, and how it all really works (and breaks) in real life. There are analytical yet not overburdened texts. And no course sales or generic words.
Interesting posts worth reading:
• About the new role of GenAI Application Engineer
• His article for Forbes — how AI helps scientists find effective drugs faster.
• What makes truly successful founders special?
• An important element of AI consulting
• Why now is a unique time to start a business
Follow — @maxvotek
Max is an entrepreneur, investor, and co-founder of Customertimes.
For 17 years, he has been implementing technologies into the business processes of large companies and living alongside alligators in sunny Florida.
Max offers honest reflections on AI, business, and how it all really works (and breaks) in real life. There are analytical yet not overburdened texts. And no course sales or generic words.
Interesting posts worth reading:
• About the new role of GenAI Application Engineer
• His article for Forbes — how AI helps scientists find effective drugs faster.
• What makes truly successful founders special?
• An important element of AI consulting
• Why now is a unique time to start a business
Follow — @maxvotek
Has anyone tested Showrunneryet?
Brief overview – this is an AI platform where you can prompt, choose the genre and details, and then generate fully animated scenes, including characters, dialogues, and all visuals. It's all aimed at creating animated series.
They opened access to the platform a year ago, but back then you couldn't generate anything yourself. Instead, a range of cartoons was available, and it was positioned as "AI cartoon streaming, where you can influence the plot" (attaching a screenshot from those times).
Now things are more serious – there's a platform for generation, but access is still through a waitlist.
#videoGenerative@TochkiNadAI
#fun@TochkiNadAI
Brief overview – this is an AI platform where you can prompt, choose the genre and details, and then generate fully animated scenes, including characters, dialogues, and all visuals. It's all aimed at creating animated series.
They opened access to the platform a year ago, but back then you couldn't generate anything yourself. Instead, a range of cartoons was available, and it was positioned as "AI cartoon streaming, where you can influence the plot" (attaching a screenshot from those times).
Now things are more serious – there's a platform for generation, but access is still through a waitlist.
#videoGenerative@TochkiNadAI
#fun@TochkiNadAI
Updated Qwen3
And Alibaba is clearly aiming to seriously compete with the other models in the category of the mostwonderful naming.
They rolled out a fresh update of their flagship Qwen3-235B-A22B.
There are many parameters (235B), but only 22B are utilized per pass through the architecture, hence the name.
The model has become faster, tests were conducted on RTX4090. In benchmarks, it surpasses Claude 4 and DeepSeek V3. In some tests, it is close behind Gemini 2.5 Pro. The context window is also increased, allowing more files to be loaded.
In essence, Qwen3 is currently one of the most technologically advanced and open models on the market, but as we know, leadership can sometimes change even daily.
Hugging Face Try online
Oh, and Qwen3-Coder is also out.
A separate version for coders. A total of 480B architecture, 35B active. A context of 256,000 tokens, extendable to 1M. For development tasks, it remains on par with Claude Sonnet and GPT-4.1, while DeepSeek V3 lags behind.
And Alibaba is clearly aiming to seriously compete with the other models in the category of the most
They rolled out a fresh update of their flagship Qwen3-235B-A22B.
There are many parameters (235B), but only 22B are utilized per pass through the architecture, hence the name.
The model has become faster, tests were conducted on RTX4090. In benchmarks, it surpasses Claude 4 and DeepSeek V3. In some tests, it is close behind Gemini 2.5 Pro. The context window is also increased, allowing more files to be loaded.
In essence, Qwen3 is currently one of the most technologically advanced and open models on the market, but as we know, leadership can sometimes change even daily.
Hugging Face Try online
Oh, and Qwen3-Coder is also out.
A separate version for coders. A total of 480B architecture, 35B active. A context of 256,000 tokens, extendable to 1M. For development tasks, it remains on par with Claude Sonnet and GPT-4.1, while DeepSeek V3 lags behind.
One person with neural networks can now do the work of an entire team. Imagine you have an AI system that every day on its own attracts an audience, runs a blog, creates content, and even sells!
🔥From idea to implementation, Denis Ffring will show the entire process of creating a system step by step at the practical workshop on July 24.
You'll see how to practically integrate neural networks into your project (or start a new one from scratch) to achieve real results. No useless prompts or drawing cats!
Live broadcast you will learn:
– How to hire an entire AI staff team that will handle the routine tasks and PROMOTE, create content, and sell ON THEIR OWN
– How to create tons of engaging non-trivial content: from posts to videos literally in a few minutes
– How to automate Telegram, Threads, Reels, VK, Zen
– How Denis' team gained 110,000 followers in a prohibited social network and 16,000 on Telegram using AI: case study.
– And most importantly, how to stop doing everything yourself and delegate up to 90% of the routine to autonomous AI agents
⤵️The workshop will be held on July 24 at 19:00 Moscow time, sign up via the link: [REGISTER]
🔥From idea to implementation, Denis Ffring will show the entire process of creating a system step by step at the practical workshop on July 24.
You'll see how to practically integrate neural networks into your project (or start a new one from scratch) to achieve real results. No useless prompts or drawing cats!
Live broadcast you will learn:
– How to hire an entire AI staff team that will handle the routine tasks and PROMOTE, create content, and sell ON THEIR OWN
– How to create tons of engaging non-trivial content: from posts to videos literally in a few minutes
– How to automate Telegram, Threads, Reels, VK, Zen
– How Denis' team gained 110,000 followers in a prohibited social network and 16,000 on Telegram using AI: case study.
– And most importantly, how to stop doing everything yourself and delegate up to 90% of the routine to autonomous AI agents
⤵️The workshop will be held on July 24 at 19:00 Moscow time, sign up via the link: [REGISTER]
❤1
This media is not supported in your browser
VIEW IN TELEGRAM
How to Create a Native iOS App with Kiro in a Few Hours
I came across an example of using AWS Kiro, a new AI-IDE with AI agents and MCP servers.
The author of the video JeredBlu created a fully functional Swift app for iOS, integrated with Apple Health and Apple Watch, in just a few hours. He wanted to manually add missed workouts if Apple Watch didn't count them.
The process:
First, he described the task in Kiro using spec-driven development mode, forming detailed technical specifications and a plan. Then, through the MCP servers, AI agents gathered the necessary documentation, looked for real cases, and broke the task into separate steps. The design, task list, and then the code with tests and Apple Health integration were automatically generated. When bugs or errors occurred, AI agents quickly fixed them, learning from mistakes through embedded memory. Within a couple of hours, a fully operational application was created.
Kiro + MCP 🤌
Full video
Waitlist Kiro
#coding@TochkiNadAI
#agents@TochkiNadAI
I came across an example of using AWS Kiro, a new AI-IDE with AI agents and MCP servers.
The author of the video JeredBlu created a fully functional Swift app for iOS, integrated with Apple Health and Apple Watch, in just a few hours. He wanted to manually add missed workouts if Apple Watch didn't count them.
The process:
First, he described the task in Kiro using spec-driven development mode, forming detailed technical specifications and a plan. Then, through the MCP servers, AI agents gathered the necessary documentation, looked for real cases, and broke the task into separate steps. The design, task list, and then the code with tests and Apple Health integration were automatically generated. When bugs or errors occurred, AI agents quickly fixed them, learning from mistakes through embedded memory. Within a couple of hours, a fully operational application was created.
Kiro + MCP 🤌
Full video
Waitlist Kiro
#coding@TochkiNadAI
#agents@TochkiNadAI
This media is not supported in your browser
VIEW IN TELEGRAM
Came across a video.
Nowadays, many creators make fake ads, and it's great, as it’s easier for viewers to compare, evaluate, and draw conclusions through familiar images.
Nicely done. Considering how conservative car brands can be regarding permissible creativity (we worked a lot with car brands, there's usually little creative freedom), this creative looks like Mercedes could have approved such a scenario. Moreover, you can see that AI is not just to reduce the filming budget by 9475948 times.
By the way, Dobrokotov recently made a cool post on the topic, and I agree with a quote from it:
Overall, the work is good; it’s a win-win, as they say. The brand gets an unusual creative for much less money, the viewer gets a somewhat restrained but not the most banal story, and the creator gets their money.
Nowadays, many creators make fake ads, and it's great, as it’s easier for viewers to compare, evaluate, and draw conclusions through familiar images.
Nicely done. Considering how conservative car brands can be regarding permissible creativity (we worked a lot with car brands, there's usually little creative freedom), this creative looks like Mercedes could have approved such a scenario. Moreover, you can see that AI is not just to reduce the filming budget by 9475948 times.
By the way, Dobrokotov recently made a cool post on the topic, and I agree with a quote from it:
ColLeAgues, with neural networks you can show what you would never have been able to show within your production budget: incredible images, new worlds, amazing plots. Stop being mediocre—amaze and be amazed.
Overall, the work is good; it’s a win-win, as they say. The brand gets an unusual creative for much less money, the viewer gets a somewhat restrained but not the most banal story, and the creator gets their money.
By the way, share your work!
What have you been working on lately: music videos, ads, reels.
And should we make such a collection of readers' work again?
What have you been working on lately: music videos, ads, reels.
And should we make such a collection of readers' work again?
If you were suddenly looking for it – here they brought Veo 3 with improved realism<ph2>
By the way, OpenArt recently went viral on X and Product Hunt as a platform for creating comprehensive stories. I haven't tested it; if you have, share your thoughts.
#videogenerative@tochkinadai
By the way, OpenArt recently went viral on X and Product Hunt as a platform for creating comprehensive stories. I haven't tested it; if you have, share your thoughts.
#videogenerative@tochkinadai
If you suddenly ignore my recommendations, then I ask you not to do so now, because I've been reading Tips AI for a long time, and I highly recommend it to you ❤️
I recently met Igor (the author), but I've always noticed that the concept of our channels is similar, yet the topics don't overlap much. Igor tests the tools himself and shares only what truly works.
Posts from the channel that are worth reading:
• Prompting scales, but verification does not
• Why do prompts not work?
• Marketing stack from what he uses
• How to get an AI project grant using AI
• Workflow for N8N
• Prompt, so that any AI writes like a human
There's also a lot of original content, memes and neuronets that he tests himself.
Overall, I highly recommend Tips AI! Make sure to check it out!
I recently met Igor (the author), but I've always noticed that the concept of our channels is similar, yet the topics don't overlap much. Igor tests the tools himself and shares only what truly works.
Posts from the channel that are worth reading:
• Prompting scales, but verification does not
• Why do prompts not work?
• Marketing stack from what he uses
• How to get an AI project grant using AI
• Workflow for N8N
• Prompt, so that any AI writes like a human
There's also a lot of original content, memes and neuronets that he tests himself.
Overall, I highly recommend Tips AI! Make sure to check it out!
Recently, I managed to persuade GPT to write in the style of Husky (don't ask).
I would give him a poem, and he would sprinkle it with Husky's characteristic tarnish, a beautiful folkloric language, sometimes with Khrushchyovka mysticism.
In short, why am I saying this?
When did GPT learn to write poems so well??
Look, there's the language, the rhyme, the meaning, all carefully measured according to poetic sizes.
Of course, it didn't happen every time, but still. It didn’t use to be able to do this.
By the way, in terms of Husky's style, this is not the best example, but I won't show you the best, they'd ban me📺
By the way, it's ChatGPT 4o.
I would give him a poem, and he would sprinkle it with Husky's characteristic tarnish, a beautiful folkloric language, sometimes with Khrushchyovka mysticism.
In short, why am I saying this?
When did GPT learn to write poems so well??
Look, there's the language, the rhyme, the meaning, all carefully measured according to poetic sizes.
Of course, it didn't happen every time, but still. It didn’t use to be able to do this.
By the way, in terms of Husky's style, this is not the best example, but I won't show you the best, they'd ban me📺
By the way, it's ChatGPT 4o.
Today I've been working with agents in ChatGPT all day
Honestly, I expected worse. I actually like it.
What we've already done together:
1. Created a work schedule for me considering a bunch of inputs.
It browsed through Notion, checked work kanbans, strategy, personal boards, accessed Google Calendar, and... Google didn't allow it to log in, considering the GPT browser unsafe. But the agent created an .ics file for me, which I uploaded to the calendar in three clicks and ended up with an excellent recurring schedule that I minimally adjusted.
2. Set up my personal Notion.
My teamspace is already great, but I hadn't gotten around to organizing my personal Notion for a while. I let the agent in, it went to the template marketplace, came up with a decent structure based on my request, found cool templates, and filled them in. The only thing is, it doesn’t know how to work with formatting, so some things need to be tidied up manually later.
3. Registered me for a fingerboarding contest.
Honestly, I expected worse. I actually like it.
What we've already done together:
1. Created a work schedule for me considering a bunch of inputs.
It browsed through Notion, checked work kanbans, strategy, personal boards, accessed Google Calendar, and... Google didn't allow it to log in, considering the GPT browser unsafe. But the agent created an .ics file for me, which I uploaded to the calendar in three clicks and ended up with an excellent recurring schedule that I minimally adjusted.
2. Set up my personal Notion.
My teamspace is already great, but I hadn't gotten around to organizing my personal Notion for a while. I let the agent in, it went to the template marketplace, came up with a decent structure based on my request, found cool templates, and filled them in. The only thing is, it doesn’t know how to work with formatting, so some things need to be tidied up manually later.
3. Registered me for a fingerboarding contest.
Yeah, I'm not even sure what else to add 🖐
Important to consider:
– Describe your request in detail, if it's as extensive as in my two points. VERY detailed. This will help you avoid a lot of iterations, as it doesn’t work very fast.
– When it accesses its browser, it's a struggle. Forms work slowly, sometimes get blocked, and entering a password or email can be a real pain.
– It seems it can't refer to memory from your other dialogues with GPT.
Attached in the post:
1. The agent bustling in my Notion.
2. A list of connectors. What you can give the agent access to.
3. An unsuccessful login attempt in the calendar (by the way, logging in through a Google account in Notion worked).
4. Example schedule. Under each agent's message, there's a button like this, you can set up recurring tasks, it's cool.
5. The agent registering me for the competition.
How is your relationship with the agent in ChatGPT starting off?
#agents@TochkiNadAI
Important to consider:
– Describe your request in detail, if it's as extensive as in my two points. VERY detailed. This will help you avoid a lot of iterations, as it doesn’t work very fast.
– When it accesses its browser, it's a struggle. Forms work slowly, sometimes get blocked, and entering a password or email can be a real pain.
– It seems it can't refer to memory from your other dialogues with GPT.
Attached in the post:
1. The agent bustling in my Notion.
2. A list of connectors. What you can give the agent access to.
3. An unsuccessful login attempt in the calendar (by the way, logging in through a Google account in Notion worked).
4. Example schedule. Under each agent's message, there's a button like this, you can set up recurring tasks, it's cool.
5. The agent registering me for the competition.
How is your relationship with the agent in ChatGPT starting off?
#agents@TochkiNadAI