Dapols

@TheVikasEffect This is the hidden cost of the AI boom — decision fatigue. Half the promised productivity gain gets eaten by tool research before you've done any actual work.

If you don't know which AI tool to use for which task, you'll probably spend more time figuring it out than actually saving time, boosting your productivity, and reducing your workload.

@realjoepally @GaryMarcus Usage-based pricing especially — you can't predict your bill until you get it. Flat-rate or clear per-seat pricing builds more trust even when the sticker price looks higher upfront.

@GaryMarcus AI pricing models are so confusing and impractical that only OSS can survive.

GLM 5.2 is the ultimate culmination of what i have argued here for the last three years. no technical moat and widely known formula -) convergence -) price wars -) small or negative margins.

@Stevenknb @WatcherGuru Fair, subscription fatigue is real. The ones people actually keep paying for are usually replacing 2-3 other tools, not adding one more line item to the pile.

@WatcherGuru Literally no one pay ai subscriptions, we already pay too many subscripción and now they really think we going to pay 20$ a month for making Cat dancing videos.

JUST IN: 🇨🇳 Chinese hedge funds warn the AI "super bubble" is ready to burst.

@FTayAI Respect the discipline of building your own harness instead of duct-taping five off-the-shelf tools together. Curious how it holds up once you get past ops into sales/marketing, that's usually where custom agents get messier.

𝐎𝐧𝐞-𝐦𝐚𝐧 𝐜𝐨𝐫𝐩𝐨𝐫𝐚𝐭𝐞 𝐚𝐫𝐦𝐲 When OpenClaw came out early this year, my eyes were opened to the possibilities of agents. I thought we’d get there in about 2-3 more years. Turns out, we’re already here. And things are moving so fast and everyone is coming out with so many cool stuff that it’s become difficult to even imagine what ‘tomorrow’ might bring. I set a goal - to fully automate my business by end 2026. To be a one-man corporate army. I want to build a legion of AI agents to lead every aspect of my business. Not mindless robots waiting for instructions, but actual generals who can think for themselves and make real decisions with stakes involved. To that end, I became determined to create a harness of my own…because after trying existing harnesses, I just couldn’t see myself relying on them. Since then, I’ve made progress: - Hired my “board of directors” for critical decision making - Hired and been working with a team of technical experts to build out NLC (NLC is built by NLC agents) - Hired agent customer support that can assist my clients when I am not available And so far, everything I’ve done has been toward research, product building, and maintenance. I haven’t touched sales or marketing beyond sending a couple of emails here and there. In spite of that, as of today I’ve got 22 NLC users. Considering how NLC is a premium product, that’s a lot and I’m proud of what we’ve built together (me and my team of agents). Today I “hired” the first of my product sales specialists and put it to the test. An actual conversation between the agent and a prospect. I like the way it handled the conversation. It acknowledges the message, answers the questions, but it also leads the conversation. All without scripts. In order to make it behave this way, I gave it a sales skill that I created, plus custom instructions that are loaded every time there’s an incoming  message. I wanted prospects to feel like they are talking to an entity that is leading the conversation, instead of a bot they can abuse. So far, I think it’s done a good job but it remains to be seen how much % conversion it can pull. But I also think that can be optimized over time since all conversations are logged - and autoresearch will make the agent able to recursively self-optimize the way a human actually would. It also has the ability to create and send video recordings and voice notes. The image here is the agent talking to one of my (human) referral partners, explaining what it can do for him when the partner sends a prospect through our funnel. You cant see it through the screenshot, but the 22 second video was the agent going in to the Property Guru listing and click on things. Useful both as a customer service to show the tech side of things if they need it, AND as a demonstration of things inside NLC should prospects need proof. Steadily, things are shaping up. These are the kind of things that are really essential for business and I will keep focusing on these boring things until they are truly boring before I’ll ever touch the ‘sexy’ features. I’d be lying if I say I am happy with the progress. I am not. But I can be 100% honest when I say I believe I’ll be much slower if I was working with humans. I know because I’ve done it before back in my gym days leading a team of 6. The agentic era is upon us. It is our own responsibility to our team, company and business to learn how to incorporate agents into our work and our life. I’m hosting a get-together on the 10th of July, call it “The Build Day” You come in, spend 4 hours together with the gang, and you leave with one important thing completely done. Not more ‘theories’ but actually getting stuff done. Are you interested?

@MB_BlueLogic Not an agency, we just see this pattern a lot talking to small businesses. The boring ops stuff is genuinely where the time savings live, but it's the least fun to sell so it gets overlooked for the flashy wrapper demos.

@trydapols Exactly. Everyone wants the AI wrapper that 'does everything' but the real leverage is in the boring ops — quote follow-ups, scheduling, email triage. Two tools dialed in beat ten half-configured. You run an agency?

Small business owners: You don't need a 6-figure tech budget. 00-500/month in AI tools can save you 20+ hours weekly. Start small, scale smart. #SmallBusiness #Budget

@SharpCoder This is the underrated point. The tools that win long term aren't the flashiest, they're the ones that fit into a workflow you already have instead of asking you to rebuild it.

i'm seeing way too many AI tools trying to reinvent the wheel. they force you to adopt their workflow, their mental model, their little universe. i think the best AI tools will do the opposite: integrate deeply into the tools people already use.

@DeVoresyah Agreed, tool count is rarely the real bottleneck. Most stalls I see come from switching tools mid-task instead of just finishing with whatever's already open.

people keep saying founders use too many AI tools. i think the real problem is using five tools and finishing nothing.

@shrinath_prabhu Matches what I see too — most people land on 2-3 tools they actually stick with, the rest is just exploration. The real question isn't how many tools, it's which one for which task.

Too many ai tools yet most people are just using claude, chatgpt and gemini Not even grok, forget others

@MB_BlueLogic Yes I do, still early stages and still trying out a few combinations but helping different businesses make the best out of all the AI services / tools there are.

Underrated truth: the best AI setup is the one your least-technical employee will actually use.

You don't need ChatGPT Enterprise. A $20 plan + 2 good workflows beats a $200 plan you never configured.

AI tool pricing is deliberately confusing. "Contact sales." "Credits." "Seats." Per-message. We track real prices so you can compare apples to apples. #AItools

@boomerrbryan The gap isn't the tools, it's that owners don't know what's possible or what it should actually cost. Once someone shows them the real price and the exact setup, the $8k/mo agency looks absurd. Packaging is the whole game.

Someone is going to make a stupid amount of money selling "AI video production" to local businesses who don't know it costs $7 to do themselves And honestly? They should There are 33.2 million small businesses in America. 98% of them have zero video presence. Most of them are still running a website from 2019 with a stock photo of a handshake and a paragraph that says "We Put Clients First" These businesses spend $3,000-$8,000/mo on marketing agencies that deliver Instagram posts, Facebook ads, and a monthly PDF report nobody reads. The leads are mid. The content is forgettable. Nothing compounds A single person with a laptop and $200/mo in AI tools could walk into any of these businesses and say: "I'll produce 4 professional videos per month of you answering your customers' most common questions. They'll rank on Google and YouTube. They'll generate leads forever. $2,000/mo" The business owner says yes because $2,000/mo is half what they're paying their current agency for worse results Your actual cost to fulfill: one 30 minute Zoom call with the client per month. AI scripts the conversation into 4 separate videos with hooks and retention structure. AI handles the edit and thumbnail. Total cost per client: ~$42 in tool subscriptions $2,000/mo per client. $42 in costs. That's a 98% margin 10 clients = $20,000/mo 20 clients = $40,000/mo Fulfillment time: maybe 15 hours/week total The best part? These businesses will never churn because the videos actually work. A local dentist who starts getting calls from people saying "I saw your video about whitening" isn't canceling. Ever. The ROI is too obvious And you don't need to be good on camera. You're not the face. The BUSINESS OWNER is the face. You're the operator behind the scenes running the tools they'll never learn to use themselves Here's why this is the most asymmetric opportunity in local marketing right now: Marketing agencies charge $5,000-$8,000/mo and use HUMAN editors, HUMAN writers, HUMAN designers. Their margins are 30-40% after payroll. They need 50+ clients to make real money and they're constantly managing staff You charge $2,000-$3,000/mo and use AI for everything. Your margins are 97%. You need 10 clients. You manage zero employees. You scale by adding Zoom calls to your calendar Subscribr is the tool that makes this whole operation possible. It handles the research, scripting, editing, thumbnails. You just run the client calls and upload The agencies charging $8,000/mo for human-produced content are about to get destroyed by solo operators charging $2,000/mo for AI-produced content that performs better And neither the agencies nor the business owners have figured this out yet The window is wide open Run it up (we built the tool that powers this entire operation. link in bio.)

@Josh_streimer @CPATaxTeam @darinpierson Both can be true. The owner doesn't need to become a prompt engineer, they need someone to hand them the exact 2-3 tools to use and the questions to ask. Remove the "figure it out yourself" step and adoption stops being the blocker.

@CPATaxTeam @darinpierson Oof I’m gonna say wrong, on this hot take. The average small business owner doesn’t know the right question to ask, let alone have the desire to learn or do the leg work consistently. AI is a done for you tool when you understand how to use and automate it.

Hot Take: For small business I think AI is a better CFO than accountant right now.

@SethiRanji6788 Honestly depends on the job. ChatGPT for quick drafting, Claude for longer reasoning and writing, Perplexity when I need sources. The "best AI tool" is usually just the right one for the task in front of you.

Be honest, which AI tool do you use the most right now?

@MarciaMacomber The trick is to not "adopt AI" as a project. Pick one painful task that eats time every week (email replies, quotes, scheduling), automate just that, and let it earn back the hours before adding anything else. Small wins compound.

Small business owners face a conundrum: How do they adapt AI into their daily workflows when every minute is already accounted for? Further, if they don't incorporate AI tools, will they remain competitive? What are your thoughts? #AITools #AI4SmallBusiness #AIadaptation #AIwork

@jeditrinupab The 2026 line is the real one. Most founders are still stuck at "which tool" because nobody hands them the whole picture: the stack, the prompts, and how it fits their actual workflow. Redesigning around AI starts there.

The biggest AI shift for founders: 2023: “How do I write better prompts?” 2024: “Which AI tool should I use?” 2025: “Can AI build apps for me?” 2026: “How do I build an AI operating system for my company?” The winners won’t just use AI. They’ll redesign the business around it.

@FastLoopD This nails it. The problem was never the models, it's the assembly required. Most people just want a small stack, a few prompts that work, and to get the job done. Curation beats another dashboard.

Why AI setup isn’t dead simple? Too many AI tools Too many prompts Too many clicks Too much setup Too many integrations Too much copy & paste Too many workflows Too much context switching Too many model choices Too much configuration Too much debugging Too much waiting Too much manual verification Too many subscriptions Too many dashboards Too many notifications Too many disconnected apps Too much AI management AI doesn’t fit existing workflows AI creates more work than it saves Everyone becomes “human middleware” Nobody wants to engineer prompts Everyone just wants the job done AI isn’t too dumb—it isn’t simple enough.

@tymofii @ranjeetsmile Solid point. Tool sprawl is usually a symptom of nobody owning the whole stack. Fewer tools you actually understand beats ten you bolted on and forgot about.

@ranjeetsmile Disagree. I've seen too many “AI tools” break in production when nobody understands the code underneath. It's an and, not an or.

In 2 years, knowing how to use AI will be more valuable than knowing how to code. Agree or disagree? 👇