aniwasbored

shopify apps charge per image to do alt text. i built the broke founder version: airtable = image database claude code = workflow setup codex = alt text generation local sync script = pushes it to shopify 240 image alt texts updated without touching them one by one.

caught claude code doing something that matters. asked it to use playwright and audit my site like a user. it looked like it was scrolling, but it was skipping chunks of the page. if you're using agents for UX audits, make them prove contiguous screenshots. otherwise you're auditing the gaps.

Someone needs to build a skill that plugs into your codebase and helps you create loops that are worth keeping

@InduTripat82427 The compute speed is nowhere close.

AMD just made Nvidia's AI strategy look very expensive. Lisa Su walked on stage with a tiny PC. Then ran a 235B parameter AI model locally. No cloud. No data center. No $4,000 GPU. Just a $1,499 machine. The secret isn't raw compute. It's memory. AMD's new Ryzen AI Max+ 395 lets the CPU and GPU share up to 128GB of unified memory. That changes everything. Most consumer GPUs hit a wall long before giant models fit into memory. This box doesn't. While everyone is chasing bigger GPUs, AMD quietly built something that can run models previously reserved for serious AI servers. And the results are turning heads: → 235B parameter models running locally → Massive memory advantage over flagship gaming GPUs → DeepSeek R1 inference outperforming hardware that costs significantly more But the bigger story isn't the benchmark. It's what happens next. Today, power users are stacking subscriptions: $200/month for Claude $200/month for ChatGPT $20/month for Cursor $20/month for Gemini Thousands of dollars every year just to access AI. Now imagine owning the hardware once and running powerful models on your desk. Private. Offline. Unlimited. No rate limits. No API bills. No waiting for cloud capacity. The AI race is shifting from "who has access" to "who can run it locally." And the people who spot that shift early usually win big. The AI hardware war just entered a very interesting new phase.

@nikks_techie Not building anything currently, does it still count?

Looking to connect with people on X who are into: • Backend Engineering • System Design • Java • Agentic AI • LLMs • Distributed Systems No growth hacks. No AI influencers. Just engineers who enjoy solving hard problems and sharing what they learn. Say hi 👋

@lumendriada I can assure you it def is

either GPT-5.6 is coming, or OpenAI is having compute shortages again. this is not the same GPT-5.5 i've been using for the last month.

So you’re telling me there are chinese nationals with american passports who get to access fable 5?

the more i use cli tools + agents, the more the split makes sense: claude code = workflows, architecture, direction codex = eyes, hands, execution one designs the machine. the other climbs inside and starts tightening bolts.

what are some actual use case for openclaw in marketing?

@marclou would this help with 72 resting?

One of my lowest Resting Heart Rate tonight. What doesn’t impact (so far): ❌ last meal before bed ❌ C02 level ❌ magnesium ❌ watching movie What does impact: ✅ Night walk ✅ Reading just before bed ✅ Dim lights ✅ Temperature (impact+++) n=500 ish nights

@steipete that makes sense. so the real advantage is sovereignty + model choice? i’m curious where openclaw wins in daily dev work vs claude code beyond the platform philosophy.

@ani3am depends if you care about an open platform and model choice or being locked in to one company that decides which prompts are okay and which will be blocked or routed to weaker models.

Wanted to buy a new Mac Studio for San Francisco. We don’t even know how to make these anymore.

@elonmusk who else feels uncanny valley with these video generation models?

Try the latest Grok Imagine!

@Rasmic @DarioAmodei they already require age checks and ID verification likely very soon

hey big head @DarioAmodei when's fable coming back? :)

@henokcrypto ASI*

Fable 5 is AGI Anyone saying otherwise has no clue what they’re talking about

@rezoundous always been this way ever since claude was launched

Anyone feeling AI getting cult-ish?

@DanielSmidstrup because @AnthropicAI realized early that the real juice is in business-facing development, not the attention economy. images/video are distribution-native. claude is workflow-native. different game.

Why doesn’t Claude have an image or video generation model yet?

@rezoundous gemini is impressive because it somehow found the overlap between “not what i wanted” and “why did this cost that much”

feeling opus 4.8 or gpt 5.5 isn't good? go use gemini for 5 minutes.

@sflorimm if you are an aspiring founder, audience first unmatched power in learning from others you become unstoppable

as a solo founder, what do you build first? - product - audience

the harder question is capital formation. you can move the entity to a more sovereign jurisdiction, maybe the caribbean, singapore, uae, etc. but frontier ai is not a normal software business. it needs talent density, compute access, institutional capital, and regulatory credibility. so the tradeoff is: more sovereignty, less capital gravity more capital gravity, more political exposure not obvious there’s a clean answer.

why doesn't anthropic just move to canada so they don't have to deal w/ all this bs

@TTrimoreau i think both are true. ai takes old-world jobs away. but it also creates the transition into the new world. the people willing to face that transition, learn the tools, and rebuild their workflow get to be part of what comes next.

Are you agree with that ? AI will create more jobs, not less

trying to get this shopify page out of yellow. LCP p75 is 3357ms. INP is good. CLS is literally 0. so it feels like the page is stable and responsive once loaded, but the first big visual is lagging. shopify speed people, what are you checking first?