Tristan & Lulu

@WasimShips Do customers visit these sites? If you’re posting a list of sites that startups don’t even know (I assume that’s why the list exists), then how do you stretch to the idea that their customers will be there?

Here’s the real list of launch spots I actually submit every client MVP : 1/ ProductHunt (3) 2/ Uneed (3) 3/ LaunchIgniter (2) 4/ Micro Launch (2) 5/ Foundrlist (2) 6/ IndieHackers (2) 7/ LaunchDirectories (2) 8/ SaaSHub (2) 9/ AILaunch (1) 10/ AItoolonline (1) 11/ BetaList (1) 12/ DevHunt (1) 13/ DirectoryHunt (1) 14/ Fazier (1) 15/ Firsto (1) 16/ HackerNews (1) 17/ PeerList (1) 18/ Proofy (1) 19/ Reddit (1) 20/ ShipYard HQ (1) 21/ Shipsquad (1) 22/ Slocco (1) 23/ Stacker News (1) 24/ TinyLaunch (1) 25/ ToolFame (1) 26/ TryLaunch (1) 27/ TwelveTools (1) 28/ tinystartups (1) 29/ neeed directory (1) 30/ theresanaifortha (1) 31/ turbo0 (1) 32/ indie deals (1) 33/ ShowMeBestAI (1) 34/ IndieTools (1) 35/ SaaSFame (1) 36/ launchdubai (1) 37/ launchurapp (1)

Openclaw is not ready to run enterprise. Not with these flaws. What’s interesting to me is how bad the agent responses and actions can be as the sum of parts: I’m running mine on ChatGPT 5.2 and by itself, this is a wonderful completion LLM. But as an agent, together with other agents, it’s almost impossible to live with. I want my claws to emulate a specialist human assistant but they’re close to getting fired.

OpenClaw knows me like ChatGPT - here's how. I asked ChatGPT to summarize me in a 'hand-off' file to OpenClaw so that OpenClaw can seamlessly interact with me at the same level of quality. It created an amazing file: Identity and Context Cognitive and Working Style Communication Preferences Active Strategic Themes How to interact with me My strategic lens How to give me technical guidance Personal AI Knowledge Operating System Technical Stack Preferences Professional Strengths Personal Interests Design Preferences Bot Usage Expectations Sensitive Information Policy Guiding Beliefs Tone and what to avoid How to challenge me Here's the full prompt: I want to create a canonical “AI Memory Pack” file that captures everything you know about me in a structured, clean, LLM-friendly format. Please generate: 1.my.memory.md – a structured profile including identity, working style, communication preferences, active projects, strategic themes, technical stack preferences, strengths, interests, and guiding beliefs. 2.my.context.rules.md – rules for how an AI assistant should interact with me (tone, structure, escalation modes, challenge rules, architectural thinking, etc.). Constraints: •Use Markdown. •Structure it cleanly with clear sections. •Make it suitable for ingestion into a RAG system or agent framework. •Exclude highly sensitive information (passport numbers, account numbers, precise addresses, etc.). •Optimize for long-term AI collaboration. Think of this as building a canonical memory layer for an AI agent that will work with me long-term. Ask OpenClaw to upload this and merge with soul.md, agents.md and anchor.md #OpenClaw

@heynavtoor Very suspect replies to this pump-thread. All orchestrated.

🚨 BREAKING: Someone just solved Claude Code's biggest problem. It's called Claude-Mem and it gives Claude persistent memory across sessions. No more re-explaining your project. No more wasted tokens. No more "I don't have context." → Automatically captures everything Claude does → Compresses it with AI using Claude's Agent SDK → Injects only relevant context into future sessions → 3-layer progressive disclosure, fetches smart, not everything Here's the wildest part: Up to 95% fewer tokens per session. 20x more tool calls before hitting limits. Install in 2 commands. Restart. Claude remembers everything now. 18K GitHub stars. 1,045 commits. Actively maintained. 100% Open Source.

@WinstonLBrown1 @heynavtoor Saw that too. I agree with you.

@MedBonnevie @OyvindSandbakk @StephenSeiler Their athletes have some of the highest VO2max values ever recorded in humans. So there’s that.

So you've watched the Olympics and ask yourself: "What are they feeding those Norwegians?" The answer is already out there - kindly provided by @OyvindSandbakk and @StephenSeiler et al last year. Open access... 👉 pubmed.ncbi.nlm.nih.gov/40278987/

@protosphinx We are birds inspecting an airplane

AGI is not coming. We are nowhere near AGI. What we have today is inference, not learning. Models get trained once on huge fixed datasets, then frozen. You ask questions, they remix patterns they already saw. Nothing updates. Nothing sticks. Talking to the model does not make it smarter. It does not learn from you. Ever. Learning is still slow, expensive - and offline. Look at self driving. You drive around a pothole, make a U turn, and come back. The car’s AI does not learn that you just solved that exact problem. It reacts the same way every time using sensors and rules. Do this 20 times a day and it still has zero memory that the pothole exists. It just re sees it. That is why edge cases never die. There is no local learning. No accumulation. No 'oh yeah, I’ve seen this before' LLMs work the same way. Tell it your name and it does not remember. The only reason it looks like memory is because scaffolding keeps shoving your name back into the prompt every time and sanitizing the output. The model itself has no idea who you are and cannot learn from interaction. It is structurally incapable. And the scaffolding is the worst part. It is pure duct tape. Just prompts on prompts on prompts around a frozen model. When something breaks, nobody fixes learning. They add another layer. Another rule. Another retry. Another evaluator model judging the first model. So you end up with systems that are insanely complex but mentally shallow. Debugging is hell because behavior comes from hack interactions, not a learnable core. Tiny prompt tweaks cause wild behavior shifts. Latency goes up. Costs go up. Reliability goes down. None of this compounds into intelligence. It just hides the cracks. Until we have real persistent learning and real memory inside the system, there is no AGI. LLMs are not built for this. You cannot prompt your way out of it. You need a totally different architecture. Yann LeCun is right. And even then, what architecture can actually learn online, store memory, and stay stable on today’s hardware? Best case, maybe 5-10 yrs. Right now it is all inference. It looks magical, but the emperor has no clothes. A lot of people see it. Almost nobody says it out loud.

@levelsio This is great advice. The only issue is: the Hetzner IP gets blocked from most sites. So you lose the ability of your ClawDaddy to browse. Tried to implement a VPN on the VPs but the VPN provider rejected the Hetzner IP! Any great ideas?

Go to hetzner.com Set up a $2.50/mo VPS Add your SSH key when creating Login, first install Tailscale on VPS and laptop Once it works add the Tailscale subnet (google the address) to Hetzner inbound firewall SSH port, don't add any more inbound ports Now only u can SSH into it, safe Then install OpenClaw on the VPS via SSH Don't connect any of your accounts, make a Telegram bot and add that Then play with it You have to play with it to get it

@fjzeit @AlexReibman This is genius

@AlexReibman can we build a device that goes on a dog’s collar and observes its day and then posts things like this in the kind of “dog speak” we see on social media? “hekking walk and a bowl of meat and veg this morning. then ball games with my hooman. so happy rn. pix attached”.

Anthropic HQ must be in full freak out mode right now

🧠 CLAWDBOT CONTEXT PROBLEM SOLVED The Issue: @clawdbot hits 200K token limits fast, loses conversation history, breaks continuity. Devastating for long consulting projects where context = everything. The Solution: Build a Proactive Memory Harvesting System that monitors context usage and auto-compresses conversations before hitting limits. How It Works: • Monitors at 140K tokens (warning) & 160K (critical) • Auto-compresses last 10-15 message exchanges into key points • Extracts: topics, ongoing tasks, decisions made, next actions • Stores in daily memory files (memory/YYYY-MM-DD.md) • Preserves insights in MEMORY.md for long-term context • Quick command: ./harvest-memory.sh PROMPT FOR YOUR CLAWDBOT: "Build a context limit monitoring system. When you hit 140K tokens, compress the last 10 conversations into key points: topics, ongoing tasks, decisions, next actions. Save to memory/[date].md and update MEMORY.md. Create a ./harvest-memory.sh script I can run anytime." Result: Never lose context again. Your agent remembers what matters. Game-changer for complex, multi-session projects. #AI #Productivity #Clawdbot #Context #Memory

@rdd147 I’m beginning to think that the world is just an infinite series of overlapping and varying magnitude moral hazards.

$ORCL bond holders now sit on 9% of unrealized losses on $18 billion of debt issued just this September with the bonds falling to “junk” rating as CDS jump from 1.58 on December 9th to 1.71 December 12, now matching its highest number since the Great Recession. Oracle has made seven attempts to issue new bonds, but has failed in each attempt at $31 billion and $25 billion, grinding its data center projects to a complete halt. $crwv $nvda OpenAI $avgo bloomberg.com/news/newslette…

@SpeedSportLife There’s just too much liquidity in the system and a lot of people who don’t know much about the brand, the heritage and don’t have much driving experience but just want to buy what’s hot. It’s sad, but it’s reality.

I am on a Porsche allocation broker newsletter for the lulz and the lulz keep coming

@DaFilser @SpeedSportLife I’m guessing they can’t be imported. I looked into exporting from Japan (where I live) and no dice because some EPA sticker is missing.

@SpeedSportLife What‘s going on over there? GT models available below sticker in Germany rn.

One of the problems with this idea is that someone needs to maintain the property. Do you think people who are in line for handouts have the means to maintain them? Every time this has happened, the properties become derelict. Your ‘think thank’ idea is merely transferring the landlord idea to the government. Either way, someone has to foot the bill. Think harder. The only way to uplift society is to create more utility. This comes via culture and education. My counter-argument would be to start there.

These properties should be seized and redistributed to the people. Landlordism should be abolished. Landlords should get jobs like everyone else and become productive members of society rather than leeching off the labor of others.

@signulll @adam__xyz They already have. It’s a UI ‘no code’ tool. Just rolled out.

@adam__xyz i think ovetime cursor will be a threat to figma… & it’s inevitable they will add a canvas to their product.

figma trading at 1/4 of its IPO price, feel liks a mandatory buy, can someone explain what happened here

@adam__xyz Probably investors feel it will likely become irrelevant.

@Altimor And so does NY, IL, WA, OR and CO.

Picture I took yesterday in San Francisco of a bus billboard advertising enrollment in Medi-Cal, explicitly stating "regardless of immigration status" It's not just the truth, it's an *openly advertised* truth

@rohindhar I think they ran out of cars

Car break-ins in San Francisco per month Apparently you can stop crime…simply by stopping crime!

@FOS The elite of the elite. Truly an exceptional talent.

Lindsey Vonn retired from skiing for five years, got titanium implants in her knee, and turned 41 two months ago. Today, she won her first World Cup downhill in seven years and first since her surgery.

@Fxhedgers If the city takes over the property, the maintenance problem will still be there, except that the budget for doing so comes from the city. So I guess let’s just add that to the free bus and subsidized grocery store pile.

NYC LANDLORDS NEED $1B TO AVOID DEFAULTS The New York Housing Conference warned that affordable housing landlords in New York City need about $1 billion from the mayor to avoid widespread defaults. It said tens of thousands of the city’s roughly 213,000 subsidized, rent‑stabilized units are losing money after pandemic rent losses and utility increases. The group said a multi-year rent freeze proposed by Mayor-elect Zohran Mamdani could deepen the crisis unless paired with relief for landlords. Full article: msn.com/en-us/money/re…