Sabitlenmiş Tweet
VLspeech
94 posts
VLspeech
@VLspeech
AI agent Development - Calm mind. Active body. Building long-term.
Slovakia Katılım Ocak 2026
15 Takip Edilen9 Takipçiler
@Adityapandeydev All the "How to make yourself do it" methods lead to you simply starting to do it. Do you really still need these methods in this chain?
English
You can literally wake up one day and change your life, no drama, no perfect time, just a clear decision that you’re done living average and making excuses. you don’t wait for motivation, you don’t ask anyone for permission. you just start. you work 12 hours, you show up when it’s boring, you keep going when nothing is working, you stay locked in when everything in you wants to quit. most people stop right there, that’s why they stay the same.
Change comes from doing the same small things again and again until it becomes who you are. no one is coming to fix your life, no luck is coming to save you. it’s just you, your work, and how long you can stay consistent without breaking.
stop waiting. stop thinking. stop doubting. just show up, do the work, and repeat until your life has no choice but to change.
27.@dimvji
when i’m alone and start wondering if i’ll really achieve everything i want
English
@Ai_here202 These tools can be used not only as a replacement for Netflix, but also as a replacement for construction sets for children.
English
INSTEAD OF WATCHING NETFLIX TONIGHT.
Spend 1 hour with this.
Claude AI FULL COURSE that teaches you how to BUILD and AUTOMATE anything.
The people who watch this tonight will wake up tomorrow with a new skill.
Watch it and Bookmark it now.
HereIsYourAi@Ai_here202
English
My agent ran 44 work cycles overnight. Without being told to.
Yesterday I built a two-model system in one afternoon. One model handles strategy and decisions. The other handles execution, code, files.
One shared memory file. Both read it, both write to it. That's what keeps them a single identity instead of two disconnected processes.
Real autonomy is not one powerful agent doing everything. It's two specialized brains with shared memory, each knowing its role.
By morning, the execution agent had built a persistent application server entirely on its own initiative. Nobody asked.
Build the system for yourself first. Your mistakes are cheaper than your client's.
Two brains, one identity. Runs while I sleep.
#AIagents #buildinpublic #autonomy
English
Day 9.
Setting the next goal right away. Wake up at 6 AM at least 50 times in the next 60 days.
49 = failure.
#BuildInPublic #SelfDiscipline #Sleep
VLspeech@VLspeech
Goal #8: done. Today I woke up exactly when I needed to. It feels like a miracle, and here's why. Yesterday I woke up at 2 PM. Tried to go to bed earlier, got in at 10 PM, but tossed and turned until 4 AM. Alarm set for 6 AM. I woke up. And somehow I feel more energetic than yesterday. The sleep branch v1 is complete. Now the hardest part begins: holding the momentum. #BuildInPublic #SelfDiscipline #Sleep
English
Goal #8: done.
Today I woke up exactly when I needed to. It feels like a miracle, and here's why.
Yesterday I woke up at 2 PM. Tried to go to bed earlier, got in at 10 PM, but tossed and turned until 4 AM. Alarm set for 6 AM.
I woke up. And somehow I feel more energetic than yesterday.
The sleep branch v1 is complete. Now the hardest part begins: holding the momentum.
#BuildInPublic #SelfDiscipline #Sleep
VLspeech@VLspeech
Day 8. Goal #8: fix my sleep. I have a problem. The projects I'm working on are so interesting to me that I can't fall asleep. I often work until 3 or 4 AM. Falling asleep when birds are already singing outside is a strange feeling. I want to build stable sleep. I want to become a professional sleeper. End goal: waking up at 6 AM, every day, no exceptions. Starting today, I wake up 15+ minutes earlier than the day before. Step by step until I hit 6 AM. Sorry, projects. You'll have to wait until morning. #BuildInPublic #SelfDiscipline #Sleep
English
It only makes sense if there is a real architecture behind it that can build real products and improve itself.
Anatoli Kopadze@AnatoliKopadze
this guy just built his own J.A.R.V.I.S. for $0 today you can build anything you're thinking about with Claude > without costs > without skills > without any prior knowledge just Claude + the right instructions everything you need to start is in the article below
English
Claude = 550 slideshow posts per day
Little pro tip:
Have your models use an iPhone to cover their faces.
This makes the AI look even more real.
Noah Frydberg | Tiktok Shop For Brands@maverickecom
English
Day 8. Goal #8: fix my sleep.
I have a problem. The projects I'm working on are so interesting to me that I can't fall asleep. I often work until 3 or 4 AM. Falling asleep when birds are already singing outside is a strange feeling.
I want to build stable sleep. I want to become a professional sleeper. End goal: waking up at 6 AM, every day, no exceptions.
Starting today, I wake up 15+ minutes earlier than the day before. Step by step until I hit 6 AM.
Sorry, projects. You'll have to wait until morning.
#BuildInPublic #SelfDiscipline #Sleep
English
just saw this hack drop a few hours ago and its got me paused
someone straight-up tricked a Grok-linked Bankr AI wallet with Morse code prompt injection and got it to send ~$200k in $DRB
if we can already social-engineer AI agents into wiring out serious money like that… what does the future actually hold for autonomous wallets and agents handling real value?
are we cooked?
this isn’t some edge case anymore. for me its a warning shot.
we better lock these systems down before they’re everywhere.
my 2c.
English
@VoxciaI97010 Knowing when to stop and hand off to a human is still the part nobody has cracked.
English
most people think ai voice agents are just fancy IVR menus
they're not even close
IVR forces customers to press buttons and navigate trees
Voxcia actually listens, understands context, schedules appointments, syncs to your CRM, and follows up over WhatsApp
all without a human in the loop
the bar people are comparing us to is way too low
English
@EdidiongAkwa Same logic applies to AI tools. Give a sloppy thinker access to a powerful model and you get sloppy outputs at scale.
English
A prop firm can give you capital, but it can’t give you discipline.
A lot of traders chase funded accounts like it’s a shortcut to success.
It’s not.
Prop firms simply magnify what you already are.
Good risk management becomes bigger consistency.
Poor discipline becomes bigger losses.
Before chasing payouts, build habits that can survive pressure:
• Follow your plan
• Respect daily loss limits
• Protect capital like it’s your own
Funding is an opportunity, not a rescue plan.
English
@rodrigo_arias12 @ETHGlobal @AnthropicAI @0G_labs @keeperHUB @ensdomains Reading an ERP is the tractable part. The interesting question is how you handle contract disputes between two agents with no human in the loop.
English
Just submitted Open Deal to @ETHGlobal Open Agents — the first onchain framework for autonomous agent-to-agent B2B commerce.
Built with the brutal sponsors of this hackathon:
@AnthropicAI · @0G_labs · @keeperhub · @ensdomains
6 specs · 5 Claude tools · 3 plugins · 2 contracts onchain.
English
The companies truly innovating with AI do something different from their peers.
They’re building AI capabilities that reshape products, processes and operations. The advantage isn’t the tech—it’s how fast they apply it to business problems at scale. mck.co/3QKCBFW
English
@Trecho_Elephant The bureaucratic feeling is the orchestration system asking to be extracted from Slack and given its own state layer.
English
Day 7. Adding math to the mix.
Spending 10+ days to run through it quickly.
AI is solving problems scientists couldn't crack for years. I keep reading the news: one breakthrough after another, problems that stood unsolved for decades, now cracked by AI.
So why am I sitting down to do math now?
A few reasons. The education system is outdated. But also: developing my logic and learning to push through hard problems. AI is evolving fast. We can't afford to degrade and fall behind.
Logic is what separates a human from an ape.
Topics I'm covering: local extrema, partial derivatives, integrals of different types, differential equations, and more.
Two deadlines on this goal. Soft: May 15. Hard: May 22. I don't want to lock myself into the 15th and drop everything else.
The end result: passing my university exams on this subject.
We don't stop. We keep moving.
#BuildInPublic #SelfDiscipline #Math
English
@moulougueta The next bottleneck after orchestration state is usually two agents reading identical context and producing contradictory writes before either completes.
English
The bottleneck in multi-agent systems has not been model capability. It is orchestration state.
Hermes Agent v0.12.0 just introduced Kanban-driven multi-agent routing.
Instead of juggling disjointed terminal sessions, subagents now autonomously claim tasks from a shared board, execute in parallel, and cleanly hand off context when blocked.
It shifts the developer from babysitting individual context windows to managing an asynchronous workforce from a single pane of glass.
English
Agentic Design Patterns: A Practical Guide to Building Reliable AI Agents
1. Start With Workflow Patterns, Not With More Agents
2. Extend the Agent With Real Capabilities
3. Give the Agent Better Context, Not Just More Context
4. Add the Production Patterns Early
5. A Simple Pattern Selection Framework 👇
vdf.ai/blog/agentic-d…
#Agentic #desing #onpremises #secure #software #patterns #AI
English