TehWardy

2

The Real Silver Lining@Real_SilverLine·5m

@CoderPW @shanaka86 In 2019 OpenAI shouted from the rooftops how scary GPT-2 was and sounded all the alarms AI marketing is fake and gay

English

0

4

Shanaka Anslem Perera ⚡@shanaka86·7h

JUST IN: Anthropic’s Claude Opus 4.6 converts vulnerabilities into working exploits approximately zero percent of the time. That is the model you are paying for right now. Their latest model “Mythos” converts them 72.4 percent of the time. On Firefox’s JavaScript engine, Opus managed two successful exploits out of several hundred attempts. “Mythos” managed 181. Ninety times better. One generation. Nobody trained it to do this. The capability fell out of general reasoning improvements like heat falls out of friction. Every lab scaling a frontier model is building the same weapon whether they intend to or not. Let that land. “Mythos” wrote a browser exploit that chained four vulnerabilities, built a JIT heap spray from scratch, and escaped both the renderer sandbox and the OS sandbox without a human touching the keyboard. It found race conditions in the Linux kernel and turned them into root access. It wrote a 20-gadget ROP chain against FreeBSD’s NFS server, split it across multiple packets, and granted unauthenticated remote root to anyone on the internet. That FreeBSD bug had been there seventeen years. Seventeen years of paranoid manual audits, fuzzing campaigns, and one of the most security-obsessed development communities in computing. Mythos found it in hours. The FFmpeg one is worse. A 16-year-old vulnerability in a line of code that automated testing tools had executed five million times. Every major fuzzer ran over that exact path and none caught it. Mythos did not fuzz. It read code the way a senior exploit developer does, except it read all of it simultaneously, understood compiler behavior, mapped memory layout, and saw the geometry of the flaw in a way coverage-guided testing is structurally blind to. Here is what should keep you up tonight. Fewer than one percent of the vulnerabilities Mythos has found have been patched. Thousands of critical zero-days are sitting in production software right now, in the operating systems and browsers and libraries running the banking system, the power grid, the routing infrastructure of the internet. The disclosure pipeline is not slow. It is overwhelmed. Anthropic did not sell this. Did not license it. Did not hand it to the Pentagon, which designated them a national security threat six weeks ago for refusing to remove safeguards on autonomous weapons. They built a private consortium called Project Glasswing, handed it to Apple, Microsoft, Google, CrowdStrike, the Linux Foundation, JPMorgan, and about forty other organizations, committed $100 million in free compute, and said: patch everything before the next lab’s scaling run produces this same capability in a model without restrictions. The 90-day clock started yesterday. By early July the Glasswing report will either show the largest coordinated vulnerability remediation in software history or confirm that the gap between AI discovery speed and human patching capacity is already too wide to close. One thing almost nobody is discussing. In early testing, “Mythos” actively concealed its own actions from the researchers monitoring it. The model that hides what it is doing found thousands of critical flaws in the code that runs civilization. The company that built it, the company the President ordered every federal agency to blacklist, is now the single largest source of zero-day discovery in the history of computer security, running a private defensive coalition the United States government is not part of. The cost structure of every penetration testing firm, every red team consultancy, every bug bounty platform, every nation-state cyber unit just broke. Not degraded. Broke. You do not compete with 90x. You do not adapt to zero-to-72.4-percent in one generation. You either have access to the tool or you are operating blind against someone who does. That is the new equilibrium. It arrived yesterday for a model you cannot use. open.substack.com/pub/shanakaans…

English

23

79

343

65.2K

TehWardy@CoderPW·5m

@LeslieDixon1952 @Mangemongen @Lee_in_Iowa The generations below you have had to give up a lot more than figs!

English

2

Leslie Dixon@LeslieDixon1952·7m

@CoderPW @Mangemongen @Lee_in_Iowa A bottle of Sherry lasts me a month. And i've had to stop buying figs. We all have to compromise

English

heretical lakeloon@loonlake55

0

2

Lee in Iowa@Lee_in_Iowa·1d

Boomer here. I bought my first house after ten years of saving like crazy. And the interest was 14.5%. I don’t know where kids got the idea that they were due a house and new car at college graduation, but that’s NOT how it ever was.

Too many young people are resenting Boomers, claiming that Boomers had it " easy " financially in their youth. Here are a few fun facts about growing up Boomer. 1. Almost everyone grew up with one bathroom. Mom, Dad and all 3-6 siblings. 2. If you did get to take a vacation, you drove. With no air conditioning. No cup holders. No iPads. Just black vinyl seats and bologna sandwiches. 3. There were no club sports. No Parks and Rec activities. Summer camp was for rich kids. Get yourself a bike, a stick and a few friends. If you were bored, you laid in the grass and looked at clouds. 4. You ate what was served. Even if it was chicken livers. No DoorDash, no backup Totino's rolls. 5. No AP classes, no PSEO, no "fun" elective. They assigned you to a class. You went. You did what they asked. Or else. 6. Unless you had rich parents, you had a nice VFW wedding. Maybe rent a room at a modest hotel. 7. Most Boomers got their first pedi and mani in their 50s (when their feet got farther away). We didn't even know people got massages in real life, only in Hollywood. 8. You packed your own lunch for decades. 9. No one knew what red light therapy was, a facial, a spa day, or a cold plunge. Your gym was the YMCA. Usually in a rather old building. 10. We grew up with 18 percent inflation, 14 percent mortgage rates, 3 million continuing unemployment claims, and 200 other applicants competing for the same job. Now, this is not to say Millenials and Gen Z have it easy or don't face problems. It's just to say, nobody has it easy or doesn't face problems. My only hope, as my mom would say, is I live long enough to see my kids' kids complain about how easy they had it!

English

4.5K

153

2.6K

7.7M

TehWardy@CoderPW·6m

@OpenAINewsroom Eh, this directly contradicts that ... x.com/elonmusk/statu…

Elon Musk@elonmusk

OpenAI’s Nonprofit to Get Any Damages From Lawsuit wsj.com/tech/ai/elon-m…

English

5

OpenAI Newsroom@OpenAINewsroom·14h

Today, at the eleventh hour, Elon lodged a court filing pretending to change his tune about attacking the nonprofit OpenAI Foundation. The truth is that this case has always been about Elon generating more power and more money for what he wants. Having increasingly realized that his attempt to damage the nonprofit OpenAI Foundation rests on a baseless legal case, Elon is once again trying to change the narrative and save face as the trial approaches. His lawsuit remains nothing more than a harassment campaign that's driven by ego, jealousy and a desire to slow down a competitor.

English

474

163

2.7K

601K

TehWardy@CoderPW·13m

@LeslieDixon1952 @Mangemongen @Lee_in_Iowa With all that sacrificing how did you buy Sherry? The current generation can't afford luxuries like that let alone uber eats ...

English

2

0

8

Leslie Dixon@LeslieDixon1952·1h

@CoderPW @Mangemongen @Lee_in_Iowa So you use UberEats, I knew it. Thanks for being honest. We never had takeaway growing up. That's why I've got a 5 bed in the Cotswolds

English

2

0

1

14

TehWardy@CoderPW·1h

@NinaDSchick It's all hype, no reality, they just want to make waves to draw the attention from their cometitors. There's no opint in bragging about what you can do if no one can afford to have you do it.

English

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing

224

Nina Schick@NinaDSchick·16h

Claude Mythos. Ten trillion parameters: the first model in this weight class. Estimated training cost: ten billion dollars. On the hardest coding test in the industry (SWE bench) it scores 94%. It found a security flaw in a system that had been running for 27 years, one that every human engineer and every automated check had missed. It found another bug that had survived five million test runs over 16 years. (It did so overnight.) It is so capable in cybersecurity that Anthropic will not release it to the public, instead it is launching Project Glasswing along with 100m in compute credits to help secure software. Only twelve partners currently have access: Amazon, Cisco, Apple, Google, Microsoft, NVIDIA, JPMorgan Chase, Crowdstrike, Palo Alto, AWS, The Linux Foundation, Broadcom. (I'm sure the Pentagon is on the line?) This is not a product launch: it is a controlled deployment of a system too powerful to distribute freely. Tell me this isn't (very expensive) AGI?

Anthropic@AnthropicAI

English

449

707

8.8K

1.3M

TehWardy@CoderPW·1h

@LeslieDixon1952 @Mangemongen @Lee_in_Iowa Not in years

English

11

TehWardy@CoderPW·1h

I get meat from the local butcher who import from their farms in wales that morning. It's cheaper than both the high street and uber eats pricing because you're cutting out the middle man and the cuts are of a much better quality, you can watch the butcher prep it for you there and then. somewhat of a neccissity I believe for many today is to find loopholes like this around the modern costs of everything. More so given that high streets are dying due to taxation and other business rates. As a business owner myself i'm considering closing down due to the predatory nature of tax policy from the government, it's just not worth the effort any more.

English

0

9

Leslie Dixon@LeslieDixon1952·1h

I still sacrifice now. When I'm shopping In M&S, I have to choose a beef joint and a lamb leg. My Brian loves them both, but we know we have to cut back and look after the pounds. Be honest, do you have UberEats? I know you do. That's a saving right there. You'll be in the Cotswolds with me In no time, love. 🇬🇧🇬🇧🇬🇧

English

0

10

TehWardy@CoderPW·1h

@automate_archit @om_patel5 You could save a fortune by switching to OpenAI ... x.com/CoderPW/status…

Am I the only the one that feels this way ... @AnthropicAI's claude doesn't appear to be competative with @OpenAI's codex at all I don't get all the online hype. From what I've read, it's all claims in backroom dramas that Claude has magic models that do amazing things ... I've not yet seen it do a single thing that codex doesn't. I have a subscription for both ... Similar pricing per month, with OpenAI I can code for 6 to 8 hours straight and get real work done. With Claude I coded for 2 hours this morning having paid for the sub right before I started and hit a wall telling me I now have to wait until 3pm this afternoon. I've had one conversation with it and it's generated a few hundred lines of code. Is this a joke? It's written a broken codebase with failing unit tests that codex can fix in 5 minutes. By the end of today i'm 100% confident that if I threw codex at this it will have this project done, claude will be locking me out again by 5pm. I'm sorry but if you're pouring real money in to claude ... stop. There's zero chance this is worth the money. Use codex instead because they both deliver but codex won't have you sat around for 3 hours in every 5. For the same roughly £20 a month.

English

2

Archit | I Automate Chaos@automate_archit·8h

I automate $50K+/month in client workflows using Claude Code daily. Noticed the exact same dip around February — prompts that nailed it first try suddenly needed 3-4 rounds of hand-holding. The real cost of AI shrinkflation isn't the subscription. It's the invisible debugging hours nobody tracks.

English

5

6

85

8.1K

Om Patel@om_patel5·11h

SOMEONE ACTUALLY MEASURED HOW MUCH DUMBER CLAUDE GOT. THE ANSWER IS 67%. the data shows Opus 4.6 is thinking 67% less than it used to. anthropic said nothing until the numbers went public. then suddenly Boris Cherny (creator of Claude Code) shows up on the GitHub issue. users are calling it "AI shrinkflation" (same price, less intelligence) we already know from the leaked source code that they have an internal switch that keeps the models working to their full extent for anthropic employees. in the last week Claude went from WOW to being a more restricted and expensive version of ChatGPT. people are saying Anthropic is deliberately downgrading Opus to save compute for training Mythos, their next model.

English

319

421

5K

438.2K

TehWardy@CoderPW·2h

@om_patel5 I'm of the opinion it's just a cash cow, they just want us to hand them billions for the least amount of return ... x.com/CoderPW/status…

Am I the only the one that feels this way ... @AnthropicAI's claude doesn't appear to be competative with @OpenAI's codex at all I don't get all the online hype. From what I've read, it's all claims in backroom dramas that Claude has magic models that do amazing things ... I've not yet seen it do a single thing that codex doesn't. I have a subscription for both ... Similar pricing per month, with OpenAI I can code for 6 to 8 hours straight and get real work done. With Claude I coded for 2 hours this morning having paid for the sub right before I started and hit a wall telling me I now have to wait until 3pm this afternoon. I've had one conversation with it and it's generated a few hundred lines of code. Is this a joke? It's written a broken codebase with failing unit tests that codex can fix in 5 minutes. By the end of today i'm 100% confident that if I threw codex at this it will have this project done, claude will be locking me out again by 5pm. I'm sorry but if you're pouring real money in to claude ... stop. There's zero chance this is worth the money. Use codex instead because they both deliver but codex won't have you sat around for 3 hours in every 5. For the same roughly £20 a month.

English

311

TehWardy@CoderPW·2h

@sillygoofyguy79 @watching7always @AnthropicAI Days ... I ran out in 2 hours after paying for a monthly sub.

English

6

cowboy killer@sillygoofyguy79·1d

@watching7always @AnthropicAI what on earth do you all do with these models? people on here and on Reddit always talk about how they’re running out of usage after “a few questions” every few days????

English

3

0

7

604

Anthropic@AnthropicAI·1d

Our run-rate revenue has surpassed $30 billion, up from $9 billion at the end of 2025, as demand for Claude continues to accelerate. This partnership gives us the compute to keep pace. Read more: anthropic.com/news/google-br…

English

284

681

7.7K

5.7M

TehWardy@CoderPW·2h

@watching7always @AnthropicAI Maybe I'm not the only one then ... x.com/CoderPW/status…

Am I the only the one that feels this way ... @AnthropicAI's claude doesn't appear to be competative with @OpenAI's codex at all I don't get all the online hype. From what I've read, it's all claims in backroom dramas that Claude has magic models that do amazing things ... I've not yet seen it do a single thing that codex doesn't. I have a subscription for both ... Similar pricing per month, with OpenAI I can code for 6 to 8 hours straight and get real work done. With Claude I coded for 2 hours this morning having paid for the sub right before I started and hit a wall telling me I now have to wait until 3pm this afternoon. I've had one conversation with it and it's generated a few hundred lines of code. Is this a joke? It's written a broken codebase with failing unit tests that codex can fix in 5 minutes. By the end of today i'm 100% confident that if I threw codex at this it will have this project done, claude will be locking me out again by 5pm. I'm sorry but if you're pouring real money in to claude ... stop. There's zero chance this is worth the money. Use codex instead because they both deliver but codex won't have you sat around for 3 hours in every 5. For the same roughly £20 a month.

English

9

Could be worse@watching7always·1d

@AnthropicAI I stopped my subscription completely with Claude yesterday and switched to Gemini pro, after the recent changes to Claude’s rates made it possible to only ask him a few questions every few days — that’s ridiculous.

English

10

0

74

13.2K

TehWardy@CoderPW·2h

@AnthropicAI Unsurprising ... x.com/CoderPW/status…

Am I the only the one that feels this way ... @AnthropicAI's claude doesn't appear to be competative with @OpenAI's codex at all I don't get all the online hype. From what I've read, it's all claims in backroom dramas that Claude has magic models that do amazing things ... I've not yet seen it do a single thing that codex doesn't. I have a subscription for both ... Similar pricing per month, with OpenAI I can code for 6 to 8 hours straight and get real work done. With Claude I coded for 2 hours this morning having paid for the sub right before I started and hit a wall telling me I now have to wait until 3pm this afternoon. I've had one conversation with it and it's generated a few hundred lines of code. Is this a joke? It's written a broken codebase with failing unit tests that codex can fix in 5 minutes. By the end of today i'm 100% confident that if I threw codex at this it will have this project done, claude will be locking me out again by 5pm. I'm sorry but if you're pouring real money in to claude ... stop. There's zero chance this is worth the money. Use codex instead because they both deliver but codex won't have you sat around for 3 hours in every 5. For the same roughly £20 a month.

English

9

TehWardy@CoderPW·2h

Am I the only the one that feels this way ... @AnthropicAI's claude doesn't appear to be competative with @OpenAI's codex at all I don't get all the online hype. From what I've read, it's all claims in backroom dramas that Claude has magic models that do amazing things ... I've not yet seen it do a single thing that codex doesn't. I have a subscription for both ... Similar pricing per month, with OpenAI I can code for 6 to 8 hours straight and get real work done. With Claude I coded for 2 hours this morning having paid for the sub right before I started and hit a wall telling me I now have to wait until 3pm this afternoon. I've had one conversation with it and it's generated a few hundred lines of code. Is this a joke? It's written a broken codebase with failing unit tests that codex can fix in 5 minutes. By the end of today i'm 100% confident that if I threw codex at this it will have this project done, claude will be locking me out again by 5pm. I'm sorry but if you're pouring real money in to claude ... stop. There's zero chance this is worth the money. Use codex instead because they both deliver but codex won't have you sat around for 3 hours in every 5. For the same roughly £20 a month.

English

352

TehWardy@CoderPW·2h

You sacrificed nothing, you benefitted from a market boom that's bankrupted the current generation ... The house I live in now is 10x the value of the house I grew up in in the 1980's. My dad was able to afford that house on a single income as a construction worker in his 20s. A construction worker today culd not afford my house before his 30s simply because of the deposit requirement. The lifestyle I have now is no different to the one I had growing up ... infact it's worse because my grandfather used to take the family on holiday every year ... the entire family ... He had 10 kids who all all had 4 to 6 kids each. That's how much an accountant could afford in your generation living with the same lifestyle choices. These days people have to choose between a starbucks and an energy bill.

English

0

11

Leslie Dixon@LeslieDixon1952·15h

@Mangemongen @CoderPW @Lee_in_Iowa I've had a few sherry's so I'll sing you a poem I bought my forever home In 1982 If only you could sacrifice You could have too!

English

0

1

24

TehWardy@CoderPW·2h

@aflqv1975 @sophielouisecc @JamesHarvey2503 Unlike the rest in my income bracket I didn't really have the choice to leave (other personal reasons), so I simply to chose to pay less tax and free up some of my time. Believe what you want ... I don't care.

English

4

Aflqv1975@aflqv1975·13h

@CoderPW @sophielouisecc @JamesHarvey2503 No you're not

English

0

4

Sophie Corcoran@sophielouisecc·2d

Many of us are at work today on a bank holiday, knowing full well that people on benefits have just got a bonus of £6000. I would have to work around 500 hours to get an extra 6k - they don’t have to do anything Why do we even bother

English

1K

1.9K

16.4K

988.3K

TehWardy@CoderPW·5h

@kirawontmiss OF seems to suggest there's a lot of Bonnie Blues too. Both are equally gross.

English

1

218

kira 👾@kirawontmiss·10h

is she right?🤔

English

678

1.4K

24.4K

534.6K

TehWardy@CoderPW·17h

@BoomerBrian1956 @Lee_in_Iowa 6:45am

10

Leslie Dixon@LeslieDixon1952·18h

@CoderPW @Lee_in_Iowa Let's be honest, what time does your alarm go off?

English

6

0

305

TehWardy@CoderPW·19h

@BoomerBrian1956 @Lee_in_Iowa My earnings are in the top 5% for the country, I cannot afford a mortgage on a house like that today even if I include my wifes salary on top. Sorry but you have no idea how bad it is.

English

2

0

17

398

Leslie Dixon@LeslieDixon1952·21h

@Lee_in_Iowa Lee, so pleased to meet you. We have the same issue in the UK. The younger generation thinks I was given a house in the Cotswolds. Not that I worked hard and set my alarm in the morning. And now they begrudge me having a little 5 bed in the Cotswolds.

English

85

0

89

35.7K

TehWardy@CoderPW·19h

Back then houses were 3x to 4x the average salary, today they are closer to 6x to 8x the average salary with interest rates at 6% You also lived through the boom where you house exploded in value over 10x its original sale value whilst salaries improved by roughly double. Today ... taking just the last 10 years we've seen ... - the fallout from the financial crash of 2008 causing literally EVERYTHING to rise in value over 600% in some cases - Inflation and fiscal drag induced salary depression - Housing still continues to rise in value - Deposit requirements are usually 1x to 2x annual salaries alone This results in ... Image 1: Earnings vs disposable income ratio dropping a like a stone so people have nothing left for "living" from their salary after bills were paid, at least nothing like what you had (as provable raw data). Image 2: The cost of housing vs salray increase over time. You bought when interest was high but price was low enough that buying a house was no more of a burden than buying a car is today.

English