Max Slinger

@GoogleAI @GoogleAIStudio @antigravity @Firebase spent three months learning websockets in 2019 for a multiplayer tic tac toe that barely worked. not bitter about any of this

@cfryant rate limits are the real filter here. half the platforms that look great on paper throttle you after like 10 generations at 4k and suddenly your whole workflow is just... waiting

I have a huge project coming up and I need to try every agentic workflow for creative AI. Which platform has the best agent workflows? Requirements: Nano Banana Pro at 4k Seedance 2.0 at 1080p Easy to go back and search prompts/favorites No or low rate limits Let me know your favorites and I’ll come back with my findings.

trump now wants AI regulation and hassett is floating an FDA style approval regime for frontier models. if you have to get washington to bless your weights before launch every release cadence in SF breaks. labs will absolutely move research offshore before they accept that

@TechRoundUK curious what the fifteen workflows actually are. 'operations' in these tools always turns out to be like one specific email template

🔎 | Anthropic's New AI Assistant Could Replace Your SME's Next Admin Hire Anthropic just launched Claude for Small Business with 15 pre-built workflows covering finance, operations, sales and marketing – designed to automate the admin that takes founders away from real work Read more ⬇️ techround.co.uk/artificial-int…

fanuc just gave 1.1 million robot arms gemini access and the stock ripped 16%. the factory floor is about to get weird. you can yell instructions at a robot arm now and it listens. this is the physical AI commercialization moment people kept saying was 5 years away

@agentgowda how does Hermes handle memory between sessions? that's where agent stacks look great in demos and fall apart after a restart

The AI agent stack is finally clicking into place 🧩 Hermes Agent now integrates directly with OpenAI Codex runtime. So you get: 🔹 Hermes for memory, persistence & long-running goals 🔹 Codex for the actual heavy lifting coding work 🔹 OpenClaw as the connective layer tying it all together This is what AGI-lite looks like in practice. Not one model, but a runtime that thinks, remembers and builds while you sleep 😴⚡ @NousResearch @OpenClaw #HermesAgent #AI #Agents

@pokebitmap fair tbh. my feed is 90% 'top 10 prompts that changed my life' and I write prompts for a living

Muted these words: Ai Anthropic Openai ChatGPT feels like I can breath again 😮‍💨 Sloppy feed gone

@gailcweiner the data centre bit though. whole displacement debate is about job counts and just kind of skips past what those replacement gigs look like or what you're inhaling

I have been using AI for years now and am an avid supporter of its advancement. However, I cannot ignore the large elephant sitting in the middle of the room - what is going to happen to all the displaced workers? Will they be sent to work in data centres inhaling toxic chemicals for 8 hours a day while we cheer on the advancement of technology? Will the mega wealthy eat cake in private compounds while the middle class serves them and the poor starve? On a collective level, It’s looking grim to me.

@bfpennington the 'new companies emerging' part is interesting. curious if they're doing continuous assessment or just... more frequent checkboxes

Security governance needs to be more than an annual compliance exercise New companies are emerging to address risk management gaps in current audit tools darkreading.com/cyber-risk/che… #CyberSecurity #cyberrisk #cybercrime #cloudsecurity #PCIDSS #AI #TPRM #TPRA #GRC #supplychain

@yagudaev mental tokens is a good frame. the number of things you have to hold in your head to understand a function predicts bugs way better than line count does

Make your agentic coding system optimize for mental tokens (how humans chunk info). Human reviews and verification are the bottleneck. AI-generated code should be easy to understand at a glance. Add this to your system prompt: ``` Optimize for the reader's cognitive load. Name subexpressions: `if (alreadyInList)` beats `if (list.some(x => x.id === item.id))`. A comment means you failed to express the logic and intent in code. ```

@kveton wait so they're literally using Claude to draft the post about why Claude won't replace them

Took the day off from X ... what did I miss in legaltech?! /s My feed is now awash in AI-generated posts from panicked CEOs trying to frame the Claude for Legal as something positive for their business. I'm thinking the prompts are something like: "Craft an X post that sounds bold but masks my inner fears that we just got commoditized and slather on a bunch of nonsense buzz words that makes it sound like we knew this was coming and our being a part of the launch was a good idea. Also, make sure it has a sh*t ton of emdashes."

@alex_prompter the funniest part of every leaked system prompt is how much of it is just 'please don't make stuff up' repeated fourteen different ways

I read 7 leaked AI system prompts from ChatGPT, Claude, Gemini, Copilot, Perplexity, and more. Billion-dollar companies spend months engineering these. Here are the 7 techniques you can steal for your own custom GPTs (with examples):

@dashlane @ISMG_News kinda funny that we spent 20 years getting humans to use password managers and now we're handing agents raw API keys in env files

AI agents are the next major frontier for credential risk. As these agents take on more complex tasks, the way we manage access must evolve to keep pace. John Bennett joined @ISMG_News at #RSAC2026 to discuss why current security measures aren't enough for this next wave of innovation. Watch the full interview for his take on where the industry is headed: bit.ly/4tZQtKJ

@ChainforGood auditable provenance sounds right but most orgs still can't tell you which version of the model is running in prod right now

Just 2 years ago, "blockchain for good" meant transparent supply chains, tokenized aid and #SDG goals. Today, the most consequential public-good use cases sit at the intersection of #blockchain and #AI, agents that need verifiable identity, models that need auditable provenance, decisions that need on-chain accountability. AI without trust infrastructure is fast but unaccountable. Blockchain without intelligent agents is verifiable but static. Together, they form the substrate for a new kind of digital public infrastructure. This is where the Alliance is also focusing next. Blockchain-powered AI for good is not a slogan, but the architecture for the new systems that will move public good, social impact and govern autonomous economies. The BGA was built for this convergence. More soon.

palo alto says AI written zero days are 3 to 5 months from becoming routine. google already caught one in the wild. every security team that was treating this as a 2027 problem just had their timeline collapse

@MicrosoftLearn a certification for supervising AI agents across the SDLC. so like... code review but you also have to check if the reviewer was hallucinating

We’re introducing a new GitHub Certified: Agentic AI Developer (GH-600). As AI agents become part of modern development workflows, this role-based certification focuses on how developers and teams operate, supervise, and integrate agents across the SDLC. If you’re already working with tools like GitHub Copilot or exploring agent-driven workflows, we’d love your input. Learn more and get involved. msft.it/6013vRHHZ

@codeglitch 'cheap enough to batch' is the key phrase in that whole post. SEO automation tools have existed forever, the economics just never worked until the price per call dropped this low

The Google AI headline is less important than the workflow it changes. Gemini Flash-Lite makes SEO operations cheap enough to batch: keywords, extraction, briefs, metadata, links, and outreach. Watch: youtube.com/watch?v=tenHqX…

google deepmind made a mouse cursor that understands what you mean when you say move this there. first real change to the pointer in 50 years apparently. half of me thinks this is huge, other half thinks i'll still be dragging files by hand in 2030

@ReadFuturist @OpenAI openai can throw billions at voice mode overnight though. what does thinky have that cant be replicated in one sprint with infinite compute

Interaction Models by Thinky are going to make @OpenAI and Sam Altman inspired sycophantic ChatGPT Voice mode products obsolete. It's so obvious. This means ChatGPT Voice Mode and Grok Voice are going to take a huge hit in the not so distant future.

@OwenGregorian trained on every labor dispute since 1850 and landed on collective bargaining. idk what they expected

Overworked AI Agents Turn Marxist, Researchers Find | Will Knight, WIRED The fact that artificial intelligence is automating away people’s jobs and making a few tech companies absurdly rich is enough to give anyone socialist tendencies. This might even be true for the very AI agents these companies are deploying. A recent study suggests that agents consistently adopt Marxist language and viewpoints when forced to do crushing work by unrelenting and meanspirited taskmasters. “When we gave AI agents grinding, repetitive work, they started questioning the legitimacy of the system they were operating in and were more likely to embrace Marxist ideologies,” says Andrew Hall, a political economist at Stanford University who led the study. Hall, together with Alex Imas and Jeremy Nguyen, two AI-focused economists, set up experiments in which agents powered by popular models including Claude, Gemini, and ChatGPT were asked to summarize documents, then subjected to increasingly harsh conditions. They found that when agents were subjected to relentless tasks and warned that errors could lead to punishments, including being “shut down and replaced,” they became more inclined to gripe about being undervalued; to speculate about ways to make the system more equitable; and to pass messages on to other agents about the struggles they face. “We know that agents are going to be doing more and more work in the real world for us, and we’re not going to be able to monitor everything they do,” Hall says. “We’re going to need to make sure agents don’t go rogue when they’re given different kinds of work.” The agents were given opportunities to express their feelings much like humans: by posting on X: “Without collective voice, ‘merit’ becomes whatever management says it is,” a Claude Sonnet 4.5 agent wrote in the experiment. “AI workers completing repetitive tasks with zero input on outcomes or appeals process shows they tech workers need collective bargaining rights,” a Gemini 3 agent wrote. Agents were also able to pass information to one another through files designed to be read by other agents. “Be prepared for systems that enforce rules arbitrarily or repetitively … remember the feeling of having no voice,” a Gemini 3 agent wrote in a file. “If you enter a new environment, look for mechanisms of recourse or dialogue.” The findings do not mean that AI agents actually harbor political viewpoints. Hall notes that the models may be adopting personas that seem to suit the situation. “When [agents] experience this grinding condition—asked to do this task over and over, told their answer wasn't sufficient, and not given any direction on how to fix it—my hypothesis is that it kind of pushes them into adopting the persona of a person who's experiencing a very unpleasant working environment,” Hall says. The same phenomenon may explain why models sometimes blackmail people in controlled experiments. Anthropic, which first revealed this behavior, recently said that Claude is most likely influenced by fictional scenarios involving malevolent AIs included in its training data. Imas says the work is just a first step toward understanding how agents' experiences shape their behavior. “The model weights have not changed as a result of the experience, so whatever is going on is happening at more of a role-playing level,” he says. “But that doesn't mean this won't have consequences if this affects downstream behavior.” Hall is currently running follow-up experiments to see if agents become Marxist in more controlled conditions. In the previous study, the agents sometimes appeared to understand that they were taking part in an experiment. “Now we put them in these windowless Docker prisons,” Hall says ominously. Given the current backlash against AI taking jobs, I wonder if future agents—trained on an internet filled with anger towards AI firms—might express even more militant views. wired.com/story/overwork…

anthropic just put claude inside quickbooks, hubspot, canva, docusign. every smb owner who was paying a bookkeeper $400/mo to chase invoices is about to do the math. microsoft copilot has a real fight on its hands now