building personal AI assistant in stealth | posting my controversial takes on AI | @usfca
LA
- Tweets 9,526
- Following 10,450
- Followers 10,723
- Likes 22,226
Pinned Tweet
i started a newsletter: the AI signal.
one AI story every morning. the one that actually matters, with my take on what it means and what you should do about it.
free. in your inbox with your morning coffee.
theaisignal1.substack.com
3
1
16
2,438
Many are saying this
I haven’t touched Claude Code or Codex CLI in a while.
The terminal era is over imo. It's the wrong interface for coding agents. Tabs are ephemeral, but context is persistent, and managing 30 tabs is pure cognitive overhead.
I don’t really need an IDE like Cursor either. I rarely navigate the whole codebase anymore.
The new primitive is the agent, not the file.
(Codex desktop app is the best agentic UI for now. But we’re still early.)
14
2
309
76,380
The coming futures market in AI compute ft.trib.al/EQeNY56 | opinion
3
10
34
36,682
Ramanujan wasn’t skipping proofs. He was hiding his chain of thought to prevent distillation.
Hardy: “Show your work.”
Ramanujan: “Sorry, reasoning tokens aren’t exposed via the API.”
25
85
1,379
20,366
.@OpenRouter co-founder Alex Atallah on the "everyone is building the same thing" take:
"This reminds me of this tweet I saw... Everybody's building an agent loop with notifications, third-party connectors, context management, memory, sandboxes, agentic web search, an always-on agent on top."
"This product is showing up everywhere. It is showing up everywhere, but these are just the new table-stakes primitives."
"A 2005 version of that tweet would be: 'Oh, everybody's building the same thing. A database, a users table, a sign-in page, a sign-up page, a profile page, a logout page. Everything's the same.'"
"There's like a lot of differentiation really. There's table-stakes needs for AI just like there's table-stakes needs for the web."
@alexatallah @amasad
.@OpenRouter co-founder Alex Atallah, in his first podcast since Stripe acquired the company, joins @Replit co-founder Amjad Masad and a16z's Erik Torenberg on why the future of AI is independence and specialization.
In this conversation, Alex walks through how the Stripe deal unfolded, why he wasn't originally looking to sell, and why "payments and inference are going to blend together."
Pre-OpenRouter, the typical AI workflow had one model provider to choose from, and little pressure on that provider to lower prices. Now enterprises are diversifying across labs and open-weight models, and every board is asking about AI costs and benchmarks.
Amjad argues if your company depends on one AI lab, it can turn into your competitor. So Replit is building the layer that lets enterprises use any model and any cloud, without being locked into either.
Alex and Amjad are split on personal agents – Amjad runs one agent across his whole company and loves the cross-domain joins, while Alex says general agents cause you to sacrifice understanding, and argues 10 specialized chiefs of staff beats one superagent.
0:45 How the Stripe deal unfolded
5:05 Why mixing models beats one model
7:25 Forcing the labs to compete on price
8:50 Enterprises want open-weight models
10:30 Every board asks about AI every month
12:25 Why companies must own their intelligence
14:15 Replit as the independence layer
15:10 Everyone is building the same agent
16:35 Why Amjad built bring-your-own-cloud
18:10 Amjad's agent that runs his whole company
19:55 Why 10 specialized agents beat one
23:35 Machines, not humans, should specialize
27:30 Guardrails for agents talking to agents
31:10 Models training their own replacements
33:45 Most tasks don't need a frontier model
40:50 Training small models on Qwen 8B
43:25 The Rust cycle is coming for AI
45:10 Fusion models: frontier quality at half the cost
YouTube: piped.video/ekK8urKHPMQ
@alexatallah @OpenRouter @amasad @eriktorenberg
22
18
188
25,625
This is how terrible the apps support in Android foldable is most of the apps is just over sized and blown up versions of smartphones apps
The apps experience and os experience is alone a major reason iPhone duo wins
187
132
2,199
706,138
Baseten joined the @OpenAI B2B marketplace today as one of the first open-model inference providers.
The best AI companies are already running a mix of closed and open models at huge scale. We're excited to give OpenAI enterprise customers the ability to use open models powered by Baseten natively within Codex and through the Responses API.
12
11
173
63,959
Apple’s website vs. Samsung’s website
the first thing Samsung does is load its chatbot 💀
60
63
3,382
314,550
If your ChatGPT/Codex Dot doesn't save your phone calls with it by default, you can try this prompt. It instructs Dot to save the transcript and summary of calls to it's memory and my brain (my organized Obsidian vault on my iCloud).
---
After each voice call, save the available transcript and a concise summary to my Brain in the appropriate project or topic. Include the date, participants, decisions, ideas, action items, owners, deadlines, and unresolved questions. Link the summary to the transcript and any related work.
Remember the durable facts, preferences, decisions, and commitments that will help you support me later. Update existing information rather than creating duplicates, and clearly distinguish proposals from approved decisions.
Preserve my words in the transcript. If the transcript is incomplete or unavailable, label that clearly and never invent missing dialogue. Don’t retain passwords, tokens, or other secrets.
Use my established Brain destination. If you can’t identify or access it, ask once rather than choosing a new location. Confirm what was saved and link to the Brain entries; if anything failed, tell me exactly what remains unsaved.
oh also, and i don't know how i didnt include this:
IT DOESNT FUCKING SAVE CALLS
so it can't recall what it was doing, what was said or nothing
for that reason alone i fucking stopped trying it
8
1
195
38,550
How do you picture a superintelligence (if true superintelligence could ever be real)? I like a trick from Dungeons and Dragons for running villains with 25 Intelligence: don't try to out-plan the players. Instead, whatever they do or roll, play it like the big bad saw it coming.
40
7
114
13,663
David Robinson used to write safety reports at OpenAI and now says its culture is ‘broken’ theverge.com/ai-artificial-i…
11
3
28
26,690
Google is changing Gemini model access starting Oct 9:
• Free → Flash-Lite only
• AI Plus → loses Pro models, Flash only
• AI Pro & Ultra → retain access to Pro models
325
144
2,477
319,858
Testing out dots?
Step 1 to see real value: give it its own email address.
Takes less than a minute.
agentmail.to/blog/give-dot-e…
4
5
53
48,593
My personal concerns about the iPhone Duo:
1. No FaceID - I used an iPhone with TouchID in the past so I’m hoping muscle memory kicks in pretty soon
2. Camera - I’ve had a pro iPhone since the iPhone 13 Pro, so a “worse” camera will be an adjustment
3. This phone seems like it was designed for right handed people. I’m hoping iOS 27.3 adds an option to make the dock be on the left side of the screen (I am left handed)
4. No action button - I use my action button multiple times a day and it will definitely be an adjustment to not have it
76
6
495
72,554
An OpenAI employee focused on AI safety has quit the company and warned that top firms aren’t doing enough to mitigate the technology’s risks bloomberg.com/news/articles/…
8
10
41
38,760
"I believe there’s about a 50% chance we all die because of the development of smarter-than-human AI systems," writes Geoffrey Irving.
time-magazine.visitlink.me/c…
72
117
263
58,095
By his own admission, David Robinson is “something of a cliché”: an employee at a leading AI company who issues a dire warning while resigning from their job.
spr.ly/6012BGx6AY
16
6
16
24,111
During a presentation on the future of RE Engine, Capcom laid out its plans to embrace AI in ‘development workflows.’ theverge.com/games/1004418/c…
10
1
21
23,181
openai's safety chief just resigned
the loudest exit in AI this year. she told the atlantic the culture is broken, after years running safety.
three safety researchers were fired days earlier. now the chief is gone too.
the labs keep publishing safety frameworks. their own safety people keep leaving
1
310
OPENAI AI-SAFETY EMPLOYEE DAVID ROBINSON QUIT THE COMPANY, WARNING THAT LEADING AI FIRMS ARE NOT DOING ENOUGH TO MANAGE RISKS AS THE TECHNOLOGY’S CAPABILITIES GROW.
ROBINSON SAID AI COMPANIES SHOULD OPERATE MORE LIKE NUCLEAR POWER PLANTS, WITH MULTIPLE SAFETY LAYERS, REDUNDANCY AND CAREFUL PLANNING TO REDUCE THE RISK OF CATASTROPHIC FAILURES.
14
7
44
27,535