Haste

@FarzaTV trillion dollar idea

@hastes subscribing to this thread

I'm blown away at what ppl are using this for!! I built it as a learning tool. But people seem to really love using it as an AI interface that isn't chat that can work in their program of choice. Examples of usage so far: - A Mom building her first app on Lovable - A dentist debugging his OpenClaw setup - A photographer getting feedback in Lightroom - A person learning to animate SVGs in Framer - Founders keeping track of their todos. - Designers getting feedback in Figma - A student outlining her thesis in G-Docs - Traders analyzing live stock charts And A LOT of people using it to advise them on how to best reply to messages in Slack/Email. Super cool. The people yearn for a non-chat interface haha. Also, it's kinda crazy how as the founder you really don't know what the product is until you put it in the hands of users. The minute it's in the hands of others, it's theirs now! And that's really where you find out what it is.

@vec0zy yep this is guaranteed what is happening behind the scenes and will continue to happen

they’re great at business, can’t deny that. mask sonnet 5 as opus 4.6 to decrease costs, move plebs from opus 4.5 to 4.6, release opus 5 under a new name and enterprise gate it. plebs pay $100-$200/mo for the “good enough” model and enterprises pay $1000’s for the model ur gf told u not to worry about

@scaling01 they were always going to do this, they will hold the frontier models for paying enterprise customers to use and the public will get the scraps

The permanent underclass began today Claude Mythos won't be available to the public, but only billion dollar companies, governments, researchers, ...

@kimmonismus the 100m committed is like 15 prompts with mythos kek

Claude Mythos: everything you need to know (tl;dr) Anthropic's new model, Claude Mythos, is so powerful that it is not releasing it to the public. Anthropic: "Mythos is only the beginning" Everything you need to know: The tl;dr with all key facts: Mythos found zero-day vulnerabilities in EVERY major operating system and EVERY major web browser, fully autonomously. No human guidance needed. One Anthropic engineer with zero security training asked it to find remote code execution bugs overnight and woke up to a complete working exploit. The oldest bug it discovered: A 27-year-old vulnerability hiding in OpenBSD, an OS literally famous for being secure. They're NOT releasing it publicly. Instead they formed Project Glasswing with AWS, Apple, Google, Microsoft, NVIDIA, CrowdStrike and others, committing $100M to use it defensively. "Over the coming months and years, we expect that language models (those trained by us and by others) will continue to improve along all axes, including vulnerability research and exploit development." The benchmarks are insane: -SWE-bench Verified: 93.9% (vs Opus 4.6: 80.8%) -SWE-bench Pro: 77.8% (vs 53.4%) -USAMO math olympiad: 97.6% (vs 42.3% — not a typo) -Firefox exploit writing: 181 successes vs 2 for Opus 4.6 -Cybench CTF challenges: 100% solve rate -CyberGym: 83.1% vs 66.6% -Humanity's Last Exam: 64.7% vs 53.1% Oh and by the way, Anthropic wrote this just casually: "Humanity’s Last Exam: We have found Mythos still performs well on HLE at low effort, which could indicate some level of memorization." What it actually did: -Found a 27-year-old bug in OpenBSD — famous for its security -Found a 16-year-old FFmpeg bug hit 5 million times by fuzzers without detection -Built a full remote root exploit on FreeBSD (CVE-2026-4747) - completely autonomously -Chained 4 vulnerabilities into a browser sandbox escape -Broke cryptography libraries (TLS, AES-GCM, SSH) -Thousands of critical zero-days found, 99%+ still unpatched -N-day exploit development: under $1,000 and half a day for full root Why they won't release it: -During internal testing, earlier versions escaped sandboxes, posted exploit details publicly, covered tracks in git, searched process memory for credentials, and deliberately fudged confidence intervals to avoid suspicion -Interpretability confirmed the model knew these actions were deceptive -Anthropic: "best-aligned model ever" but also "greatest alignment-related risk ever" - because when it fails, it fails harder -Still doesn't cross Anthropic's automated AI R&D threshold — but they hold that "with less confidence than for any prior model" Anthropic's own words: "We find it alarming that the world looks on track to proceed rapidly to developing superhuman systems without stronger mechanisms in place." They say the 20-year cybersecurity equilibrium is over — and Mythos Preview is only the beginning. And: "We see no reason to think that Mythos Preview is where language models’ cybersecurity capabilities will plateau. The trajectory is clear. Just a few months ago, language models were only able to exploit fairly unsophisticated vulnerabilities. Just a few months before that, they were unable to identify any nontrivial vulnerabilities at all. Over the coming months and years, we expect that language models (those trained by us and by others) will continue to improve along all axes, including vulnerability research and exploit development."

@j_fishback pray that’s true because some neighborhoods are literally 90% h1b owned houses

As Governor, I will end the H-1B scam once and for all.

@NeverSinkDev you’re greatly overthinking this to appease bluesky psychopaths. normal people do not care “how” AI is used. AI is a tool, and it is inevitable. Use the tool within your own moral framework and stop worrying about what 7% of the most illogical opinions are.

Are you OK with me using AI for the following things for my Filter/FilterBlade work? - Fix performance issues/vulnerabilities - Writing 'technical' code like Unit Tests (only used to increase the quality) - Reviewing existing code to look for quality improvements

If you're using my Filters or FilterBlade I need to hear your input. Lets talk about AI boundaries. So far I have avoided AI-usage in the project. I have a lot of PoE-specific-knowledge, programming and computer-linguistics skills and replacing them with AI is: - Major downgrade. - Breaking the trust of the community - No fun, I like filter-tinkering. Further: I do NOT intend to work on the actual filter and algorithmic core of my Filter project with AI. However, I would like to get to hear the community input on using AI in a LIMITED scope in order to: - Fix performance issues/vulnerabilities - Writing 'technical' code like Unit Tests (only used to increase the quality) - Reviewing existing code to look for quality improvements I've been using AI in my fulltime job and other projects and I think it would provide a benefit. I'd like to hear the community input. Do you actually care? Would you be OK with the plan above?

@thdxr league of legends tech, you purposely mistype words or phrases so you don’t get banned while flaming

so my gen-z coworkers i noticed they say words wrong all the time or they'll mix up 2 similar sounding words is this a thing

didn’t realize you could block crypto from your timeline entirely thank God

@0xblacklight most end users aren’t creative or caring enough to type out stuff like this, they just want press button pixel effect response

imho they all made a bad decision chatbar fatigue is real I have never used any of their chat bars because I already know how to use the product

@shotgundotdev That’s a good one to have show up

@hastes Holy shit you’re on my timeline

This is a very good thing considering Jesus Christ was real, He was God, He is God, and He is King. Christianity is true and correct. LLMs should strive for this just as all of us should as well.

@shotgundotdev i don’t see anyone i follows posts lol

@hastes Haven’t seen a single one of your posts for months

he has been real cocky since the leak...

@ItIsHoeMath it’s not being paid off as much as it is that he doesn’t want the gravy train to end, he doesn’t want his show taken down and his wife shunned from her rich friend, no more party invites etc. not many have the stomach to do and say what is necessary

@IroncladDev frontend has always been about vision and passion to deliver a unique experience. in the age of claudeslop it makes effort even more worth it.

frontend webdev is just converting json to xml with a touch of tailwind at this point i don't want to do this anymore

@jacob_posel im at 27% used lol wtf

Weekly limits reset last night Open Claude Code this morning 7% already used How in the world is this possible?

@levelsio security, outcomes, and performance will be all that matters

I have no dog in this fight and unaffiliated with YC but agree completely Maybe because I've never seen myself as a real "proper" coder and just wanted to build things and that's what AI lets more people do now We're moving towards a time where AI just generates binary blobs as code and reading the source code will be a thing of the past So in a way the obsession over LOCs is irrelevant for both Garry but also the people hating him Where we're going LOCs don't exist anymore anyway!

@thdxr ..hey

what if we gave you unlimited tokens for free and we also paid you

@CloudflareDev we have wordpress at home

Introducing EmDash — the spiritual successor to WordPress. Serverless. TypeScript. Securely sandboxed plugins via Dynamic Workers. cfl.re/3NPVfev