Adrian

1.3K posts

Adrian banner
Adrian

Adrian

@AdrianBarbir

ceo @doelabs -- @zfellows @ycombinator

San Francisco, CA Katılım Kasım 2014
1K Takip Edilen1K Takipçiler
Adrian retweetledi
Eno Reyes
Eno Reyes@EnoReyes·
@petergyang @cursor_ai I actually think this is more of a lesson around business model and ICP selection. The pressure comes from bad margins on their largest customer segment (solo devs that are used to heavy subsidies) Being model agnostic is actually a superpower for coding agents when done right.
English
1
3
20
1.1K
Adrian retweetledi
Beff (e/acc)
Beff (e/acc)@beffjezos·
Overhead in SF: "TBPN is just Cocomelon for VCs"
English
60
129
2.8K
119.6K
Slash
Slash@slashapp·
$100M @ $1.4B Shoutout @RibbitCap @GoodwaterCap @khoslaventures @NEA @ycombinator @mickymalka @erickimSF @vkhosla @evamalonsor @rickyang @bradflora @jaisajnani @santinestea @lecheverriae @whrobbins @Joshuabrowder
Slash tweet media
Victor Cardenas Codriansky@victorcardenas

We (@slashapp) raised a $100M Series C at $1.4 billion valuation to build the world's most powerful business banking platform.¹ The round was led by @RibbitCapital, and co-led by @khoslaventures & @GoodwaterCap. And we're releasing Twin: the world’s first AI private banker.

English
28
22
254
35.2K
Adrian retweetledi
Han Wang
Han Wang@handotdev·
We just raised a $45M Series B at a $500M valuation led by @a16z and @SalesforceVC to build the knowledge infrastructure for AI
English
289
128
2.7K
596.6K
Adrian retweetledi
rahul
rahul@rahulgs·
for everyone trying to build ai products or whatever there's literally just one product to build. any successful llm app is literally just: llm in a for loop with tools and prompts products like these scale with intelligence, and will continue improving exponentially, so maybe don't try to innovate over this (many such cases)
English
66
34
1K
62.2K
Ivan Burazin
Ivan Burazin@ivanburazin·
I connected OpenClaw to Clickhouse, Stripe, QuickBooks, and Brex, and tried to pull some financial data for a board meeting. Asked: "How much did we spend on compute?" Got the figure. Next: "Show me all revenue and cross-reference with Stripe." Gave partial data as Stripe limits the API to 100 events per month (we have thousands) Brex sent an error: "We can only give you spend through this tool, not revenues." Even our CRM can do everything except upload files through their API. All the data is still siloed. Even with agents built into each SaaS tool, I still have to cross-reference everything manually. The entire approach is built backwards.
English
20
1
35
6K
Adrian retweetledi
Walden
Walden@walden_yan·
I see agent builders under-reacting to this and its implications around open source and Mythos. You should prepare for a future where we basically have AGI but its prohibitively slow/expensive. Agents will look more like fast/cheap models making requests to their "smart friends". We've found this to work especially well with mixing models from OpenAI/Google/Anthropic. But future models will also be trained around this. It's already the direction we're headed with our SWE-series models, which more than just being very smart themselves, will be very fast at doing work while delegating hard decisions to the smartest models.
Claude@claudeai

We're bringing the advisor strategy to the Claude Platform. Pair Opus as an advisor with Sonnet or Haiku as an executor, and get near Opus-level intelligence in your agents at a fraction of the cost.

English
16
8
305
48K
Adrian retweetledi
Jeff Huber
Jeff Huber@jeffreyhuber·
the vast majority of AI-failures today are not reasoning failures but context failures
Jeff Huber@jeffreyhuber

English
1
3
41
4.1K
Chamath Palihapitiya
Chamath Palihapitiya@chamath·
This may be a dumb question but I’ll ask it here anyways: I can’t find a good way for my various AI chats to automatically sync its conversation history into a structured knowledge base. So that as I update various chats from time to time and refine context, my knowledge base automatically grows with this new info.
English
1.1K
63
2.4K
800.1K
Adrian retweetledi
Jean-Denis Greze 💡
Jean-Denis Greze 💡@jgreze·
I drew this matrix a while back. I still think it's useful but it's also incomplete. There's a third axis missing that changes most of the answers: what does it cost your engineers not to build the other thing? When the market is moving fast like it is right now, your engineers/PMs/designers are the scarcest thing you have. They should be on the one or two things only you can build. Everything else, find a vendor, sign the shorter contract, move on. The opportunity cost of getting that wrong is higher than whatever you'd save by building it yourself. Vibe coding has made this harder to see. Building feels cheap now but it isn't. We sign vendors all the time that wouldn't survive this matrix. We sign them because the engineering time is worth more elsewhere.
Jean-Denis Greze 💡 tweet media
English
1
2
10
704
Adrian retweetledi
Aditya Agarwal
Aditya Agarwal@adityaag·
Everyone wants to avoid taking on the models head on. Surprising how few people take the opposite approach Lean into the risk. But make a bet on capabilities that are 12-18 months out. The models suck at something today but I am going to bet it is good enough after XXX months
English
18
8
74
8.9K
Adrian retweetledi
Gadi Borovich
Gadi Borovich@GadiBorovich·
There are 7 days left until we close applications for our next cohort. Some of the people you’ll be meeting: @pedroh96, founder of @brex @viccpoes, founder of @krea_ai @fernandorojo, head of @v0 Rodrigo Schmidt, first Brazilian at Meta @eliast, founder @agency @saradu, founder @andocorporation @victorcardenas, founder of @slashapp Andres Perez, founder Sanas. @APolakof, founder @useflai @RohanVasishth, founder of @getbluejay_ai @zacbakerr, founder @tryterac @abustamante, founder @bem_hq @Andydy42, founder @RespanAI @cvander, founder @platzi @pablorpalafox, founder @happyrobot If you’re in doubt, apply! The thing you’ll quickly realize in SF is that everyone is just like you, so please don’t underestimate yourself.
Gadi Borovich tweet media
English
9
24
81
18.6K
Adi Singh
Adi Singh@adisingh·
Who’s building OpenClaw for enterprises?? Been following this space closely. Think there’s a big winner that will be born here.
English
185
3
199
39.9K
Adrian retweetledi
Grace Isford
Grace Isford@graceisford·
product/ui/interface are moats that matter.....solving the human behavior loop - adapting humans to work hand in hand with agents continues to be one of the bigger bottlenecks in AI adoption 👇
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

English
0
4
19
4.3K

Keşfet

@petergyang @cursor_ai @ml_angelopoulos @slashapp @RibbitCap @GoodwaterCap @khoslaventures @NEA