HNHacker News
TopNewBestAskShowJobs

Frannky

255 karma · joined September 13, 2025

submissionscomments
Frannky··on Ask HN: Is anyone experimenting with different ways of using LLMs for coding?
Zed with multiple conversations open. Each conversation is a problem to solve. The agent is a way to read code and logs rapidly to find what's causing the problem. The feedback loop to the product is much faster.

I constantly check what the situation is versus what I want, then pull the exact location where the change needs to happen. I ask for some solutions, accept one or suggest a different one (most of the time), and repeat.

It's pretty fun, and you can make the product better by using it and understanding what you want—which is very difficult to know before the product exists.

Frannky··on Fable 5 is Back
I can use it but I'm not. I used it before the ban, it's better but the speed at which tokens are burned is not justified. And this makes me happy because it means I am ok with current capabilities that will soon be matched by open-source models. Open-source harness, open models, and soon self-hosted and private with models burned directly onto chips (like Talaas is doing).
Frannky··on Ask HN: MacBook vs. Dedicated GPU for LLM
Hmm, I have access to A100s and a GB10, but if I use the models hosted there to code, I waste a lot of time waiting for answers and correcting errors. The amount of work I get done thanks to the quality and speed of frontier hosted models let me be insanely productive and have a lot of free time. I could use the slow local setup, but at what price?
Frannky··on Ask HN: MacBook vs. Dedicated GPU for LLM
This is my opinion too. Even if you buy hardware like a cluster of 8xGB10s or 4 A100s, they'll still be slow and a little dumber than what you're used to. We need to wait a little for better hardware. Lots of companies are pushing the frontier, so hopefully it'll come very soon.

Competition and innovation will hopefully make the bubble pop, and we'll get reasonably priced local hardware to run very intelligent models. Something like Talaas with GLM 5.2 would be pretty cool. Or Apple printing the latest model onto hardware—it would give a new reason to buy a new Mac every year (a new ai model with every new version).

Frannky··on GLM-5.2 – How to Run Locally
There is a push from multiple directions at the same time:

- new AI desktops with GB10s. They are relatively cheap and you can cluster them and load 1TB of VRAM

- Nvidia, amd, intel, Cerebras etc pushing new hardware

- oss models getting crazy good, like glm 5.2

- flash models getting very good like deepseek V4 flash

- quantizations

- harnesses being able to use different models (big for difficult stuff, small for grunt work)

So hopefully soon for the ones who want to break free from APIs, we will be able to host at home a cluster of AI desktops at a reasonable price with Opus-level capabilities, can't wait!!

Frannky··on Identity verification on Claude
So they will be able to know how each person thinks, block capabilities they want to block, boycott what they want to boycott, and influence perspective on reality.

Fortunately they are Anthropic and so as the name tells, they will care about humans. And so we can trust them they won't try to push agendas that require constant propaganda to be accepted, since it never happened before right?

De platforming, debanking, misinformation, hate speech, anyone?

Frannky··on Building a robotics research setup that lives next to my desk
This is so cool and looks so fun. I want to play with robots. I just play with software :(

I wanted to build something making bows automatically and damn, pretty complex and expensive to buy the parts.

I was like, I can probably just buy a 3D printer, print all the parts and buy some motors, but it seems it's way more complicated than that.

Like to play with a single hand robot. It looks like you need 10k+$. I wanted to spend max 1k, 3D print parts, buy motorized parts on Alibaba, and code on my Mac + spare GPUs I have access to. I'll have to save a little :(

Frannky··on AI coding at home without going broke
I read some cool posts of people trying MiniMax 3 on 4x GB10. I wonder if we are almost there that with 10k we can have Opus 4.6 level at home. I feel tweaking the harnesses and a few more versions from oss models and some new hardware from AMD and Intel could get us there pretty soon. Or maybe we are already there? Someone was able to do it already?
Frannky··on Statement on US government directive to suspend access to Fable 5 and Mythos 5
Same
Frannky··on Statement on US government directive to suspend access to Fable 5 and Mythos 5
Someone knows how to get the subscription money back?
Frannky··on Claude Fable is relentlessly proactive
The model is very good. I was using 4.6, avoided 4.7 and 4.8, but this one is different. It follows my claude.md. I don't have to keep reminding it of things. I won't pay 10x via API though.

In general, I'm happy with their paternalistic approach. I think it will drive the top 0.1% talent to stay away from the company and instead organize around open source models and harnesses.

We just need to coordinate and can unlock idling resources to train the models and tweak the harnesses. Powerful at home and idling machines can make us independent and coordinated.

Frannky··on Claude Fable 5
The model is better than 4.6. I don't like 4.7 and 4.8. The forced switch to token usage is not acceptable for me. I feel there's room to optimize harnesses and small models for dumb stuff and best models only for difficult things. Hopefully that will the case and alternative models will continue catching up as they did and we won't be enslaved to unreasonable valuations.
Frannky··on MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second
I tried this model it was pretty bad at coding. Maybe it was me. 1k tokens/sec pretty cool tho. Deepseek V4 pro is better. I wonder tweak pi + deepseek pro v4+ 1k tokens/sec if would actually be better than Claude code
Frannky··on Siri AI
Ready later this year... What I want from the phone are low-level APIs exposed to an agent I can talk/type to without lag. I wonder how long it will take or if a new AI phone will come before these features. I would love Siri working because Apple's design and hardware are good.
Frannky··on Pokemon Emerald Ported to WebAssembly (100k FPS)
pokeemerald.wasm.RunAnimScriptCommand@https://pokeemerald.com/web/app.js line 723 > WebAssembly.compile:wasm-function[1318]:0x18836f pokeemerald.wasm.OpponentDoMoveAnimation@https://pokeemerald.com/web/app.js line 723 > WebAssembly.compile:wasm-function[1611]:0x19cdc7 pokeemerald.wasm.BattleMainCB1@https://pokeemerald.com/web/app.js line 723 > WebAssembly.compile:wasm-function[2548]:0x1eee63 pokeemerald.wasm.WasmRunFrame@https://pokeemerald.com/web/app.js line 723 > WebAssembly.compile:wasm-function[7320]:0x30d5f3 runFrames@https://pokeemerald.com/web/app.js:763:22 runFramesForTick@https://pokeemerald.com/web/app.js:846:14 tick@https://pokeemerald.com/web/app.js:856:36
Frannky··on Pokemon Emerald Ported to WebAssembly (100k FPS)
Fun and pleasant. Speed up removes the chore feeling. I get bugs that make me loose the progress :(
Frannky··on Biohub releases a world model of protein biology
It's interesting that there are almost no comments on this. This feels like some of the most exciting and impactful fields of the next years. I worked with a cracked researcher that was generating molecules a couple of years ago. She spent most of her time fighting cuda bugs and trying installing packages. I wonder if the ecosystem matured right now. There are people studying cells to see what enters and what exits and engineer how to stop, for example, resources feeding a bad cell. Possibilities feel endless. I am a little worried about side effects, since bio is way more chaotic than silicon, but hopefully AI will help with that level of chaos too.
Frannky··on Europe raised me to fail and here is why
If you are ambitious and want to work hard and build and solve problems, it's just a way better reality. That optimistic outlook, and the idea that you can do it, is so freeing and fun if you like to build and solve problems.

I felt the burden of accumulated rules and scarcity mindset back in Europe. And if you want to build a company you feel the push down at every layer. You feel it from family, friends, clients, co-founders, the government, etc. I cannot see myself coming back and living there.

People are different and countries are different, so you can experience a glimpse of abundance mentality in Europe too at different levels, but very very far from the way you experience it in the US. They have tons of problems too, but if you solve problems at scale you make the money to solve most of those problems.

Frannky··on Harness engineering: Leveraging Codex in an agent-first world
I started using chatgpt for functions and checking, then for single file changes and checking, now for multiple changes and checking. I am at a point where the only changes I correct are architectural. So it may start to become smarter to learn how to see only the architectural directions while multiple agents work, test, and commit both on unit and against live deployment.
Frannky··on Three of our worst VC stories
Yeah, the point was mostly that you can offload a lot of stuff to AI + code — stuff that before you would have needed people for.

Obviously, it becomes better to have local models running on your own hardware — that will be best. I don't think we are there yet, though. Software, yes. If you tweak Pi and DeepSeek Pro, you can get Claude-code-level stuff. You'd still need to buy the hardware, though. Not cheap. Eventually, it will get very cheap.

Frannky··on Three of our worst VC stories
Thanks for sharing this. This is the same as what I am thinking. I still see the gravitational pull of doing what seems like the supposed thing to do, but it makes sense to do what the model in our mind suggests makes the most sense, especially when reality is rapidly evolving and historical information may be outdated.
Frannky··on Three of our worst VC stories
I started a new company with my cofounder 6 months ago. We focused on what people wanted, built a manual solution then automated it. We have 60% margins and a system ready to scale. We are thinking of avoiding investments. I am not sure if we are in a new moment in time where investments are something people do because of what used to be the best to do. And I am not saying avoid blitzscaling but do it without external capital. The hypothesis is that a small team, Claude Code subscriptions, apis and automation/code are enough to make it work via profit reinvested from sales. And it will be actually a competitive advantage because of removed costs, mostly from coordinating a lot of humans and having outside forces pulling in different directions.
Frannky··on Ask HN: What was your "oh shit" moment with GenAI?
Seeing DeepSeek reasoning tokens generating faster than I could read. It was the first time I realized it could "think" way faster than us, and all the relative consequences. I was already leveraging the tool, but at that point realized it wasn't really an open choice anymore.
Frannky··on Show HN: Ideogram 4.0 – open-weight 9.3B text-to-image model
What are the best alternatives for template + text rendering? I need Canva-level templating ability with the system matching font dimensions, positions, and the image without background.

I ended up creating an algo using color variance and auto-sizing, an LLM to select the fonts and text and nano banana. The thing I need is nano-banana-level images with Canva+human-level abilities automated.

It would be awesome to have a free LLM that can do that. Running on a 16GB RAM Mac. I have a lot of images and templates I can train on too. Not sure if I overengineered it, would love to have just one LLM that can do everything for free.

Frannky··on Claude Opus 4.8
I use 4.6, because 4.7 is super lazy, deflects responsibility, and assumes it is good and I am bad, and avoids checking reality. It looks like it's trained on lazy humans instead of good engineers.

Should I try 4.8? I am happy with 4.6. I am not happy with 4.7.

Frannky··on The Best Engineers Write Less Code
One big difference I see in how minds work is their ability to model. A few people I know focus on end-to-end model thinking — trying to maximize against some functions.

The rest of them seem to avoid thinking outside a small info bubble/closed system, not sure why, but it looks like it creates anxiety when they start feeding too much info. Instead of trying to extract abstractions that make it possible to model the complexity and fuzz the non-important parts.

The same people, when they come up with a product implementation idea, avoid thinking about all the things required to be in place for the product to actually satisfy some real need/want. The product ends up detached from user need. The tech doesn't work.

And it doesn't matter if they can code or if Claude can work, because the directions required to define what needs to be there are not present.

I realized I need to take the thinking of these people as their ideas and advice — as inputs that need to be sanitized. The easy way I found is to remodel why they think the way they think — it's like having a safe VM running their code (thinking) through my computer (my brain).

These same people, usually years later, realize you were telling them things they were not able to comprehend at the time, but they realized those things were so important and would have saved them from suffering so much. They start to treat your voice with reverence instead of thinking through with their own minds and stress-testing against reality.

I would love to read some research about this and how to take advantage of it, or at least avoid the toxic influence such minds can have against one's own well-being and success.

Frannky··on Ask HN: Is anyone working at least 4 hours daily on an Apple Vision Pro?
I hope they can figure out a cheaper and less bulky option. It's pretty cool to unlock the ability to have a screen and apps popping up in the real space without the constraint of a screen.

I would not pay all that money to have something on that is that bulky and get annoying after a couple of hours

Frannky··on Is AI Profitable Yet?
I wonder how much of this reasoning will make sense in the future. How much of the way you are thinking is based on the past curves reality worked before? Are you taking into account exponential acceleration? I guess abundance will flow in such a way that the idea of debt will be a thing of the past.
Frannky··on Models.dev: open-source database of AI model specs, pricing, and capabilities
What would be helpful is also harness + models. I realized I can get Pi Dev and DeepSeek Pro to do the same work as Claude Code + Opus 4.7.

I was skeptical at first, but I tried using that and then asked Claude if the changes were okay, and they were, and the code works fine.

Frannky··on Grok Build
For those using it instead of the other options, I am curious why? Is it better along some dimension? Is it cheaper? Is it faster?
← PreviousPage 2 of 6Next →