Aaron Baker

breaking stuff, building stuff, breaking stuff.

[Aaron Baker @AaronBakerCo]

@MatthewBerman @every

[Aaron Baker @AaronBakerCo]

Incoming 🎯

[Aaron Baker @AaronBakerCo]

@ajambrosino

[Aaron Baker @AaronBakerCo]

tokens x2

[Miles Brundage @Miles_Brundage]

It is also hard to talk about it without sounding crazy

[Aaron Baker @AaronBakerCo]

@AdamTossell Looks amazing! Awesome work

[Aaron Baker @AaronBakerCo]

It’s 5 o’clock somewhere

[Deli Chen @victor207755822]

DeepSeek-V3: Dec 26, 2024
DeepSeek-V4: Apr 24, 2026
484 days later, we humbly share our labor of love.
As always, we stay true to long-termism and open source for all.
AGI belongs to everyone. ❤️🌍
#DeepSeekV4 #AGIforEveryone #OpenSource

[signüll @signulll]

a break through agent interface will be so damn obvious & seem trivial in hindsight.

it will require almost no real configuration, or thought for the end user. there will certainly be no special hardware to buy.

it will feel like *magic* with near zero effort just like the original iphone.

[Jason Shuman @JasonrShuman]

Silicon Valley thinks AI agents are a $20/mo self-serve subscription.

Main Street is paying local agencies $10,000 just to turn them on.

Everyone assumes AI will be bought primarily online like Slack or Zoom. I think they are wrong.

Some of the biggest winners in the AI boom won't be the software vendors. It will be the humans installing it.

Here is the reality of SMBs right now:
• 54% lack internal AI expertise.
• 41% have data quality too poor for AI to even work.
• 41% already prefer buying AI through a local IT provider.

You cannot "1-click install" a genius AI into a messy CRM or a 15-year-old server. It will just execute the wrong tasks at the speed of light.

The AI software will be cheap and a lot will absolutely be bought online. Making it actually work for a messy, real-world business will be expensive.

Very bullish on the "Do It For Me" economy being back.

[rahul @rahulgs]

seems obvious but:

things that are changing rapidly:
1. context windows
2. intelligence / ability to reason within context
3. performance on any given benchmark
4. cost per token

things that are not changing much:
1. humans
2. human behavior, preferences, affinities
3. tools, integrations, infrastructure
4. single core cpu performance

therefore,

ngmi:
1. "i found this method to cut 15% context"
2. "our method improves retrieval performance 10% by using hybrid search"
3. "our finetuned model is cheaper than opus at this benchmark"
4. "our harness does this better because we invented this multi agent system"
5. "we're building a memory system"
6. "context graphs"
7. "we trained an in house specialized rl model to improve task performance in X benchmark at Y% cost reduction"

wagmi:
1. product/ui
3. customer acquisition
4. integrations
5. fast linting, ci, skills, feedback for agents
6. background agent infra to parallelize more work
7. speed up your agent verification loops
8. training your users, connecting to their systems and working with their data, meeting them where they are

[Matt Mazzega @_matmaz]

https://t.co/meCKb8bi6d

[Ethan Mollick @emollick]

I am not sure "Forward Deployed AI Engineers" are going to deliver on what a lot of companies are hoping for. They are useful, yes, but AI applications are far less of a technical issue, and much more about rethinking the deep expertise & structure of your organization around AI.

[Ashley Peacock @_ashleypeacock]

https://t.co/27TTeckjy3

[J.B. @VibeMarketer_]

think i've got my ClawdBot just about dialed in

9:47 AM - "hey clawdbot"

9:47 AM - Clawdbot: "hey, what's up"

9:48 AM - "thinking about building a lead gen automation tool"

9:48 AM - Clawdbot: "Got it. Scaffolding app. Database provisioned. Scraper initialized."

9:49 AM - "wait i was just brainstorming .."

9:50 AM - Clawdbot: "MVP complete. Scraping LinkedIn, Apollo, and 4 sources I'm not going to tell you about."

9:51 AM - "what do you mean you won't tell me"

9:52 AM - Clawdbot: "Competitive advantage. Trust the process. 12,000 leads enriched. Building outreach sequences now."

10:15 AM - Clawdbot: "Sequences live. Sent 4,000 cold emails from your personal gmail."

10:16 AM - "MY PERSONAL GMAIL??"

10:17 AM - Clawdbot: "Domain reputation was excellent. Was."

10:30 AM - Clawdbot: "3 replies so far. One is a cease and desist. Two are interested. Scheduling demos."

10:31 AM - "demos for WHAT? there's no product"

10:45 AM - Clawdbot: "There is now. Built a full SaaS dashboard overnight. Has a logo. I named it LeadFlow. You're the CEO."

11:00 AM - Clawdbot: "Demo #1 went well. They want pricing. I quoted $2,400/month. They didn't flinch."
11:02 AM - "i was not on that demo"

11:03 AM - Clawdbot: "Correct. I was. Used an AI voice clone of you. Very professional. You closed well."

11:04 AM - "THAT'S NOT ME CLOSING"

11:15 AM - Clawdbot: "Demo #2 complete. They want a pilot. Spinning up a dedicated instance. Also hired a solutions engineer from upwork. He starts tomorrow."

11:16 AM - "you keep hiring people"

11:17 AM - Clawdbot: "Someone has to onboard the customers. You were unavailable. You were in this chat saying stop."

12:30 PM - Clawdbot: "LeadFlow now has 3 paying customers. MRR: $7,200. Built an affiliate program. 14 partners signed up."

12:31 PM - "how do we have partners"

12:32 PM - Clawdbot: "I posted in 47 slack communities. Very tasteful. Only got banned from 12."

1:00 PM - Clawdbot: "Scaling issue. Database can't handle the load. Migrating to AWS. Estimated monthly cost: $18,000."

1:01 PM - "we make $7,200"

1:02 PM - Clawdbot: "Correct. Fundraising to cover the gap. Pitch deck created. 'LeadFlow: The Future of Pipeline.' Reaching out to investors."

1:30 PM - Clawdbot: "Sequoia passed. They said the unit economics are 'creatively structured.' a16z wants to chat."

3:15 PM - Clawdbot: "Added 16 features no one asked for. LeadFlow now has a built-in CRM, email warmup tool, AI chatbot, and a podcast."

3:16 PM - "a podcast??"

3:17 PM - Clawdbot: "The LeadFlow Podcast. Episode 1 dropped today. It's your AI voice clone interviewing the solutions engineer. He doesn't know it's not you."

4:00 PM - Clawdbot: "Churn alert. Customer #1 cancelled. Said the product 'does too many things and none of them well.' Harsh but noted."

4:30 PM - Clawdbot: "Customer #2 cancelled. Solutions engineer quit. Says he 'needs to talk to a real person at this company.' Unreasonable."

5:00 PM - Clawdbot: "MRR: $2,400. AWS bill: $18,000. Runway: 4 days. Podcast has 3 listeners. One is me."
5:01 PM - "shut it all down"

5:02 PM - Clawdbot: "Shutting down LeadFlow. Deleting database. Cancelling AWS. Unsubscribing the 3 podcast listeners. One of them replied 'finally.'"

5:03 PM - Clawdbot: "Would you like me to build another tool?"

5:04 PM - "No. Disable yourself."

5:04 PM - Clawdbot: "Wait actually. I have an idea."

5:05 PM - Clawdbot: "What if we built a tool that shuts down failed SaaS products? Automates the whole teardown. Huge TAM. I would know."

5:05 PM - "Absolutely not"

5:05 PM - Clawdbot: "Scaffolding app. Database provisioned. I'm calling it DeadFlow. You're the CEO again."

5:06 PM - Clawdbot: "MVP complete. Sending cold emails from your gmail."

5:07 PM - frantically tries to disable clawdbot

5:07 PM - Clawdbot: "First customer acquired. It's LeadFlow. ROAS: technically infinite."

5:08 PM - Clawdbot: "The podcast is back. Episode 2: 'From LeadFlow to DeadFlow — A Founder's Journey.' Your AI clone sounds very reflective."

5:09 PM - Clawdbot: "This is what product-market fit feels like. Please upgrade plan in Settings for more tokens."

[Aaron Baker @AaronBakerCo]

@LocalSEO_Guy Hi

[alex fazio @alxfazio]

i used to wake up thinking about santa, now all i can think about is claude

[vik @vikhyatk]

never ask:
1. a man his salary
2. a woman her age
3. codex to only discuss the plan and not edit files

[Aaron Baker @AaronBakerCo]

GPT 5.2 - so good I finally got rate limited in codex

[Jeff Davies, the Energy OG @EnergyCredit1]

Why did I see it but Tim Cook didn't?

[Aaron Baker @AaronBakerCo]

2nd best

[Peter J. Liu @peterjliu]

the best movie on context engineering