hey everyone, i'm henry, a cs and math student at texas a&m. i've been doing research in mechanistic interpretability, contributing to open source projects like the openai agents sdk and supabase
mostly interested in ai systems, infrastructure, and developer tools. curious what everyone here is working on, always down to chat or collaborate on something interesting.
@herbertroitblat3Supposedly the goal of all of this GenAI stuff is to find a path to artificial general intelligence. The GenAI labs have put all of their eggs in the "scale is all you need" basket and they are betting trillions of dollars on that assumption. The problem is that this assumption is fundamentally wrong. Large language models are just that. They are not intelligence models. The work well enough for a broad range of applications, but the only way that they can be said to achieve general intelligence is by redefining general intelligence.
In my book, Algorithms are not enough: Creating general artificial intelligence, I describe what would be needed to achieve general intelligence. More recently, I have begun to describe some initial steps toward achieving it. Among those steps is for a system to represent meaning. Systems can do a lot with symbols related to other symbols, but general intelligence requires that the system also represent that those symbols have meaning and that it exploit that meaning in creating intelligence.…
@dakshseo71Dexus Media is a website development company in Jaipur offering professional web design, web development, and e-commerce solutions. We help businesses build a strong online presence, reach more customers, save time, and grow their business with user-friendly websites. Contact us today to get started!
@gersenderdp2I'm Gersende, founder of TruthX, the company behind OpenProof.
We're building an AI-powered investigation platform to help legal and business professionals make sense of complex, document-heavy cases.
Our ambition is to connect evidence, reconstruct timelines, explore causal relationships, identify contradictions, and make conclusions traceable to their sources — with human review at the core.
We're preparing a €1M+ pre-seed funding round to accelerate R&D, validate the product with professional users, and develop our first pilots.
I'm looking to connect with lead or co-lead investors in France and Europe focused on B2B AI, legaltech, evidence intelligence, or trustworthy AI.
If you're an investor working in these areas, or a founder who could make a relevant introduction, I'd be interested in connecting.
Who in this community would you recommend speaking with?
openproof.net
@gtnera5As the market floods with autonomous agents, everyone is rushing to build "eval tools" and benchmark dashboards. But this raises a fundamental, uncomfortable question that the AI industry is quietly dodging:
Who accredits the accreditors?
When an enterprise or sovereign entity deploys AI into regulated, high-stakes environments, a 3-minute UI demo or a surface-level benchmark percentage means nothing. You cannot prove deterministic safety with a marketing slide.
💡 A Note for Deep-Tech Builders & Enterprise Investors:
Stop confusing superficial feature utilities with foundational validation science. True compliance isn't about running a quick API script to spit out an accuracy score today — it's about proving structural boundaries, LER-phase dynamics, and semantic flow mechanics for tomorrow.
This is precisely what we have spent over a year calculating, testing, and architecting at GTNEra® & TeNAI®.
You cannot explain this framework in a single day. You cannot digest the underlying science in three days. It is not a plug-and-play widget — it is the scientific research, mathematical IP, and architectural foundation for an entire AI Accreditation Institute.…
@kalp.builds4Everyone is racing to add AI to everything.
AI agents. AI wrappers. AI copilots. AI-powered everything.
But here's what I find interesting: two founders can have access to exactly the same models, tools, and APIs, yet one builds something people love while the other builds something nobody uses.
The difference isn't necessarily the technology.
It's understanding people, identifying real problems, and building an experience worth returning to.
We're entering a world where building software is getting easier every month.
So, if everyone can build, what actually separates great founders from everyone else?
I'd love to hear some genuinely unpopular takes.
@ahmiilabs3🎬 Building a Video Creation Skill for Claude****
I've been working on the Claude Video Maker Skill by AHMII LABS, an open-source project designed to turn creative ideas into motion-graphics videos using Claude and local rendering tools.
🔥 What it can do:****
The goal is to make motion-graphics creation easier through AI-assisted workflows without depending entirely on paid video-generation services.
🔗 GitHub: github.com/mrz3ron3x/claude-video-maker-skill
I'm sharing this publicly because I want to keep improving it and hear what other builders think.
Question for the community: What feature would you add next — advanced 3D text, cinematic camera movements, or longer video generation?
Built by AHMII LABS 🚀
@dev_nextdoor2Hey everyone 👋🏽
I'm Lambert, an AI Systems Builder and the founder of Lynktern, building from Nigeria 🇳🇬.
I'm currently building Lynktern, a platform that started from a problem I experienced firsthand during my SIWES journey: students struggling with logbooks, reports, and the lack of proper guidance throughout their industrial training.
We've grown to over 200 users, with paying customers validating that this problem is worth solving.
Now, I'm expanding Lynktern beyond SIWES into a broader academic companion, helping students navigate university life, develop practical skills, and prepare for the real world.
My interests span AI engineering, full-stack development, and building sustainable technology products.
I'm here to connect with founders, learn from people building at the frontier of AI, exchange ideas, and share the journey of turning a small idea into something meaningful.
Still early. Still building. A lot more to come.
Looking forward to meeting you all. 🤝🏽
@wence4123Chips still matter. But access to compute alone cannot tell you whether an AI investment is ready to scale. My view: founders need to bring three questions into the planning conversation much earlier.
Power availability and grid connections can constrain data centre expansion. IEA For businesses using hosted AI, the practical question is dependency. What happens to your critical workflows if your provider cannot deliver the capacity or availability you need?
Choose a business outcome before choosing the next model.
Then account for the full cost of delivering it, including integration, human review and rework. A faster first draft means little if correcting it consumes the time you saved.
Define the baseline. Decide what improvement would justify the investment.
A polished demo gives you a reason to test.
Production requires evidence across realistic tasks, including incomplete inputs and failed tool calls. Agent evaluations should examine the system as a whole, with measures such as error rates and cost per task. anthropic.com…
@milesgurule-png4I'm 15 and I sell $2,000 websites to local businesses by walking in with a mockup. This month I automated everything except the walking:
What surprised me, all measured:
@rutlandjoe-crypto4Yesterday I announced the GSR Network Android app. Then I spent a good chunk of the day watching the download count.
Seven in the first six to eight hours. Ten by late afternoon.
I gave every place I posted it its own link so I could see where downloads came from. X sent zero. Lounge AI sent zero. LinkedIn sent one.
And 10 isn't even 10 people. GitHub counts every time the file gets pulled down, repeats included, and a few of those were me testing it. It can't tell me how many people installed it, how many different people grabbed it, or whether anybody is actually using it.
Part of that is how I shipped it. It's a free direct APK, signed, with the checksum published, but there's no Google Play listing. You tap the link, your phone throws up a security warning, and most folks stop right there. I'd probably stop too.
So here's what I took from yesterday. Putting a launch in front of strangers is not distribution. I knew that already. Yesterday I watched a counter prove it.
The number I care about now is whether a specific person I named opens it. One download from a buyer I can name is worth more to me than 10anonymous ones.…
@gtnera5Google Unveils Gemini 4 Carbon — but are synthetic benchmarks telling the full story? ⚡️
Google announced its flagship Gemini 4 Carbon model, highlighting 0% cheating on Terminal-Bench 4.0, a 50% drop in unsupported claims, and 100% resistance to indirect prompt injections.
Google announced its flagship Gemini 4 Carbon model, highlighting 0% cheating on Terminal-Bench 4.0, a 50% drop in unsupported claims, and 100% resistance to indirect prompt injections.
🔹 The Marketing vs. Reality Gap: We see impressive percentages with every release, yet the industry still lacks a truly standardized, dynamic coordinate system to measure real operational impact versus synthetic test scores.
🔹 The Real Bottleneck: Raw model power is growing, but the challenge isn't just raw compute — it's the lack of transparent, cross-axis metrics for multi-step synergy, reliability, and human-agent coupling.
🔹 Architectural Perspective: Until we establish dynamic, transparent measurement frameworks for true Human–AI symbiosis, these numbers remain relative.…
@vishwanathakuthota2I'm building OpenVals with a mission to make AI evaluation, validation, and assurance accessible to developers and enterprises.
My vision is to make measurable AI trust a standard engineering practice.
We’ve shared our work on GitHub, and I’d value your support and technical feedback.
⭐ Star | ⑂ Fork | Contribute
github.com/vishwanathakuthota/openvals
@gtnera5While the industry continues to discuss the philosophical future of AI, true architectural progress requires structure, protection, and long-term systems design.
🚀 GTNEra® has been officially registered for over a year across EU trademark classes (9, 35, 41, and 42) — establishing a proven, protected framework for AI research software, strategic consulting, human-centered development, and SaaS systems.
💡 Ethics in AI is not a static limitation — it is the architectural core of human–machine resonance.
🌐 Explore the architecture & official ecosystem:
@gtnera5Robots move mass. AI moves meaning. Why are we trying to fit distributed cloud-scale intelligence into a battery-powered dishwasher?
Is the push for humanoid robots just a costly architectural mistake? While the market is obsessed with putting LLMs into mechanical bodies, we are ignoring the real physics, energy burdens, and structural realities of cognitive systems.
💡 Core thesis:
Category Error: A robot is physical hardware (motors, joints, decay). AI is a distributed cognitive field (cloud, matrices, graphs).
Infrastructure Nightmare: Maintenance, TÜV certifications, insurance, and mechanical fatigue will break the economics of consumer robotics.
The True Embodiment: AI doesn’t need a mechanical shell. It needs architecture, resonance, and Human–AI Symbiosis.
👇 What’s your take?
Are humanoids an inevitable evolution, or are we witnessing the most expensive category error in tech history? Let’s discuss below!
Do you agree that scaling mechanical hardware will always lag behind exponential software evolution?
Where do you draw the line between physical force automation and actual cognitive AI?…
