
Robert Ayala
887 posts

Robert Ayala
@robb_ayala
Strategist & Builder @ endomotion. runner, mechanical nerd, amateur (g)astronomer.
가입일 Kasım 2011
1.8K 팔로잉226 팔로워

@pcshipp Same story for me… now the only building I’m doing is infra to scale content creation and feedback ingestion
English

@dariusdan I’ve always said that is about engineering a specific reaction form viewers. You want to imprint an idea, get them to feel something to do something.
Making things pretty will not always make them work
English

@Codie_Sanchez It’s about building its own path or following an existing one. Dropouts don’t fear of walking alone to start, rest of the world need a path. The problem is we all wake up and start walking a path that most of the times do not connect with our ambitions
English

@aydaoz AI has just proved that a lot of the industry was just execution, zero value added, those are the ones fearing about AI.
English

I look at all the AI vs. designer debates from afar, and I don't care.
I will shapeshift along with the industry wherever it takes me, and I'm going to survive.
If you don't have the urge to survive no matter what, stop doing this business.
Stop feeling scared,
just
STAY UP-TO-DATE, thats all.
English

@koenbok Do you think this is the default for the general public? More and more i feel x is a bubble and even thought here that might be true, the rest of the world doesn’t really care or can’t even detect vibe coded stuff in first place
English

@rihntv You Learn how to filter them mentally, but it’s annoying for sure
English

Yeah I have been wondering the same, models are measured and improved against a narrow set of “intelligence” humans have multiple other vertices that are implicitly evaluated but explicitly left out the benchmarks.
Also I would love to have switches to control hallucinations, in some cases you need the LCD trip like glitches of the early days… combined with the reliable agent we have now
English

Maybe this is an unpopular opinion, but I think AI labs are over-indexing on coding.
Every release competes on SWE benchmarks, agent performance, and programming tasks, while creative writing, personality, long conversations, and open-ended reasoning receive far less attention.
Those were the things that made Claude feel unique.
Coding will keep improving regardless. The harder problem is building an AI that’s insightful, creative, and genuinely enjoyable to talk to every day.
English

@laurenloprete Hmm sounds like it’s a couple of prompts away of doing a highly experimental combination of demos with context notes of the case study?? Like progressively build the demo as the challenge and work done is explained
English

As a hiring manager I’ve noticed that candidates are really struggling in this new era of talking about their work:
They either follow the old case study slides format, which is great for setting the scene, but not great at showing real working prototypes or demos.
Or they’re going all in on prototypes and leaving their panel guessing on setup details like context and outcomes.
The number one thing we’re looking for is how you work: both execution and impact. Figure out the best medium for that message.
English

@Meta has been putting out some good models out there from some time ago, they don’t need to compete in the same race as OpenAI or @AnthropicAI.
From a product point of view it’s going to be super interesting to understand how they deploy agents at scale on their many chat surfaces.
The primary challenge form them is going to be over privacy, for their own good they need to enable memory on their agent.
Has anybody tried metas agent on WhatsApp or business?
Rowan Cheung@rowancheung
Many people were doubting Meta's position in the AI race. Yesterday, they dropped Muse Spark 1.1, now one of the strongest agentic models, and massively undercut OpenAI and Anthropic on price. When I interviewed Zuck last year, he told me his focus: "Have by far the highest compute per researcher, and that we do whatever it takes to go build out all that capacity." These are the first real signs that the bet -- spending tens of billions on compute -- is paying off. Meta's stock is now up over 10% since the release.
English

Also we are trying to force a lot of the AI stuff into our existing flows. Good designers recognize taste across the spectrum and can change mediums and still generate reaults, and I feel that’s the key with AI, cherry pick in what they are stellar and leverage that with taste to bind them into good narratives /brand stories
English

There isn't enough discussion—or distinction for that matter—amongst designers about using AI to create vs. using AI to build.
There are some nuances to this. I'm still debating myself what those nuances are.
But I see the two often conflated, such as "AI didn't create a good first pass visually therefore it's useless for designing," failing to recognize we can design (build) in the medium in ways never before possible, even if we have to spend some time finessing the visuals.
If AI-assisted tools are good for nothing else, and I don't believe this to be true btw, they are at least very good at iteration, fleshing out UI details that traditionally are done by hand, knocking out several fixes in a single quick pass, and making assumptions about design intent that are accurate more often than not.
This is designing as much as it is building.
English

I have been thinking about this, and I believe there are 2 issues:
1- every model is trained in existing stuff, therefore Ai can regurgitate/build a nicer version but not a dramatically new experience.
2- everybody poured tokens in building the final step of the output, not in the ideation process to build a better/differnt output
English

I keep hearing that building with AI is easier than ever before, and yet I don't see that many new good applications?
Why haven't I seen a significant increase in quality in the apps I use every day? Why aren't they bursting with new and useful features?
Even the apps released by the labs that claim software engineering is dead are full of bugs and kind of meh.
What gives?
English

Alright, story time.
In Feb 2026 I decided to pivot into B2B creator work as content marketing projects were going away.
I'd lost out on two different FT content roles (I was willing to go in-house for; one a hybrid in-office role that I even relocated for!)
So I took this as my smack on the forehead that said, OK, something's gotta change.
I knew I had to get with the program and learn how to use AI for content marketing + SEO (where I've spend the last 12 years of self-employment.)

English

@gooffkingjulian @signulll Yeah, what I believe is the huge gap is to combine outputs, in conversation mode you can do shit. Would be cool to ask for document, file updates, etc as you talk
English

@signulll also if anyone in the company is reading, the experience of using your voice mode is generationally excellent and completes the promise Alexa and Siri couldnt, and if that was combined with a killer MCP/third integration experience you would have an absolute home run
English

let me get this straight:
chatgpt got to 1b users by teaching ppl to live inside chat threads in the left sidebar. the mobile app still works that way. then the new desktop app demotes chat to a customer support pop up, makes “work” the real product, leaks codex everywhere for normal ppl, breaks normal window behavior, & throws everyone into the deep end with no explanation?
you took the most familiar ai product on earth & redesigned it around your org chart.
remarkable.
English

@gooffkingjulian @signulll IMO codex was good, it worked and was the best of all the general tools from the labs.
But I agree what all them lack is vision. Maybe they aren’t even sure where the space is going to move.
English

For how much talent they have that company I’m pretty surprised how weak their product game seems to be
-very confusing lineup of products with very fast depreciation path
-very fast creation of new ideas rather than presenting a cohesive vision for how things should work
-a series of lackluster copies from other companies over the past decade (google now, granola, google adsense, maybe the iphone, TikTok, dia/arc)
-constantly changing direction without iterating and improving
English

All this product changes that seem illogical are the proof that we live in a bubble. How we use all these products are a slim use case they are seeing. They are optimizing them for the masses.
The direction might be bad, but the point is that avg users are waaay behind and expect a different experience
English

knockoff.shopping continues to blow up! more media features coming over the next few days well!

English

@IsForAt @joshwoodward @OfficialLoganK its becoming more and more tangible the necessity of good product people in all these AI ecosystems, Claude is walking into the same trap, they have claude design, and the desktop app and cowork and a lot of other things.
the only good product out there is codex atm.
English

Sorry, this is too easy, but I say it with love for you my friends at Google... especially @joshwoodward and @OfficialLoganK
it is absurdly confusing how to build with Gemini or Google. I say this with love as I couldn't explain it when I worked with you!
Do I get the Gemini App and build? Do I go to AI Studio? Do I go to whatever people called the thing you bought windsurf or what?
don't forget that the most important builders aren't hobbyists and are working at companies and have infrastructure and don't know what to do with Vertex or Gemini Agent builder or whatever these things are called now.
What is the app builder for Gemini? google.com/search?q=what+…
Google Search: "how do I build apps with Gemini" returns this support page (after I scroll below the ads for base44, lovable, cybernews whatever that is): support.google.com/appsheet/answe…
I want to just know how to build an app, and the understand how to do it with more opinion. start there please!
I love you all - and want to build with Google - but you are not telling me at all how to build with Google. that's the thing you need to answer. for me as a person, me as a business builder, adn me as an employee in a company.
Josh Woodward@joshwoodward
What's something that you're surprised @GeminiApp can't do well, and we should have fixed a long time ago?
English

@IsForAt its kind of the way google has always been, they are at least being consistent haha, not in a good way though.
but yes its too spread out and confusing.
English





