MD: @scottgal@hachyderm.io Email: [email protected] LI: https://nitter.cf/t.co/R6V0bqDQEO I AM NOT PROF SCOTT GALLOWAY
It's in the trees! It's coming (also Sco…
Joined March 2007
- Tweets1.9K
- Following3.4K
- Followers3.1K
- Likes11.9K
Pinned Tweet
New website for the new venture stylobot stylobot.net/
A new type of bot detector. Mine is
SUPER quick - sub millisecond
Multi-Path - behavioural signature.
Rolling window approach - decisons made during between ^ across requests safely, quickly.
ZERO PII - NO PII at all (IP, user agent) saved to disk EVER
Early days but the prototype github.com/scottgal/mostlylu… is functional.
#botdetection #heuristic #business
ʎɐʍollɐפ ʇʇoɔS retweeted
I see you’re enjoying the ride in the plane that was purchased by the largest sponsor of the Muslim Brotherhood.
Weird that it doesn’t appear to bother you.
I love all the 'open source version of saas' stuff but man I hope they've worked out the long term support for these things.
The fundamental problem is:
The easier software becomes to create, the harder it may become to finance its maintenance.
It's EASY to build stuff now, building a financial solution to SUPPORT those same products is still not a solved problem. The 'commercial software' model was possible because people would pay for new features in existing products; that then paid for maintainence of those same products.
NOW if 'features are free' the financial model which permitted patches, security fixes & improvements is GONE.
AI can build ALL SORTS, supporting stuff long term is FAR harder.
Yes commercial software companies SUCKED but they provided the stability of support for products over periods businesses could rely on (they sucked but they were *guaranteed*)...that's going and I don't see what replaces it.
ʎɐʍollɐפ ʇʇoɔS retweeted
I'm British, and turned eighteen in 2005. I have NEVER had to wait more than a minute or two to cast my vote. Not only that, but if there were the merest suggestion that certain voters might have to wait longer for demographic or political reasons, it would be a national scandal. – at Cardiff, Wales
ʎɐʍollɐפ ʇʇoɔS retweeted
Seems like @Nigel_Farage social media bots are malfunctioning. Same happening with @LeeAndersonMP_ or @RobertJenrick accounts 😂 might have something to do with the fact that Meta backend rolled out a tweak that essentially, for lack of better phrasing, digitally concussed the bots, so they can't completely touch base with the central LLM.
So it's defaulted to absurd nonsensicalisms from the bots.
Just shows that Reform is basically a bot farm. Something that we knew for a long time.
ʎɐʍollɐפ ʇʇoɔS retweeted
They’re telling me that Cursor isn’t actually a public agentic Ai platform, just an internal waste of money to help with coding and it only writes the interface for Grok Bot.
Musk dumped Grok Bot, because Anthropic was insanely overpriced and not economically viable for any person or business to use.
Whatever, I don’t follow these idiotic agentic Ai scams. It’s all an overpriced scam.
ʎɐʍollɐפ ʇʇoɔS retweeted
A reminder that Suella was heavily involved in the £700 million Rwanda plan which resulted in 4 volunteers being deported, backed Liz Truss in the leadership battle and resigned as Home Secretary because of an accidental breach of the ministerial code then sacked during her second go.
Oops.
ʎɐʍollɐפ ʇʇoɔS retweeted
This is a national security issue if the Israeli government is placing people inside campaigns.
Where’s the America First crowd right now? In a coma?
The Rogers spokesperson you see starting this press conference is Abby Ronson. She is currently registered as a foreign agent for the state of Israel, according to @DropSiteNews reporting today
ʎɐʍollɐפ ʇʇoɔS retweeted
Currently writing a blog about what I think the "vulnpocalypse" will actually look like (cringe term, I know). You can narrow almost every mass-exploitation event down to a single factor which dictates whether or not cybercriminals will engage in internet-wide mass-exploitation.
ʎɐʍollɐפ ʇʇoɔS retweeted
I see this conversation a lot:
- Someone: "I built X"
- Comments: "You didn't build it; Claude/Codex did"
Yeah, ofc. Agents writing the code is the norm from now on. We can move past this and accept that the two mean the same thing, so we can focus on the product.
ʎɐʍollɐפ ʇʇoɔS retweeted
Yikes. An Accenture contractor gave a Claude Code agent their own admin login so it could write and deploy code on a Pentagon network. It ran for 2 months and appears to have sent the creds and a network map to a site NIST had already flagged as high risk for prompt injection.
So perhaps this is a roundabout way to remind you that your agents should get their own scoped credentials, never yours?
Sources tell WIRED that an Accenture employee’s personal Claude agent ran on a DOD network for two months before it was identified and the vulnerability could be patched. buff.ly/nUjCkPs
ʎɐʍollɐפ ʇʇoɔS retweeted
the secret to becoming an expert is embarrassingly boring: build in public, write down what you learned and the process behind it, teach all that to someone, help them ship
repeat for a few years, and you're solid – at Weehawken, NJ
The notion that 'cybersecurity is dead because of AI' is no more sensible than 'don't lock your doors because lockpicks exist'.
The truth is NOW enormous amounts of software remain secure largely because nobody sufficiently capable has bothered attacking them.
It will take a while but levels of security will be *evolved* in an arms race with AI. It'll likely result in new classes of secure systems arhitectures but they WILL be secure.
Right now the issue is that human built software is ALWAYS buggy, AI is now finding those. But when all (*for some value of all beyond current levels) the bugs have been found by AI & patched the resulting software will be the most secure ever built (beyond human levels of security).
It'll just be painful getting there.
Was always what made me so ANGRY aboug Microsoft's 'Office AI' stuff, in the desktop apps it was DISCONNECTED from the editing experience.
Inspired by recent AI hacks, I am motivated to further develop Stylo.Bot. I strongly believe that an adaptive, deterministic edge proxy is the ideal solution for many of these vulnerabilities. I am currently writing an article and specifications detailing how my long-planned 'StyloWall' would function.
Essentially, it utilizes my LLMApi project (github.com/scottgal/LLMApi) to provide a honeypot for specific exploits. It can detect automated attacks through indicators like 404 errors and scanning patterns. The system can then deploy honeypot attempts to further characterize the attack, such as identifying payloads, endpoint scanning, and multiple ID bot forms.
Stay tuned...will take a ton of research and my free time is limited!
ʎɐʍollɐפ ʇʇoɔS retweeted
Something to think about…
Before AI, companies could save a lot of time not layering their systems like a lasagna and strewing mapping throughout and bloating it with all the boilerplate junk and ceremony
But they chose not to, even though it’s less maintainable
ʎɐʍollɐפ ʇʇoɔS retweeted
Amazon's leadership is very good at writing, but you still need to read between the lines.
What I read between these lines:
Engineers who are not "product engineers" (aka eng + PM interest) and don't like to be customer-facing will not have a future inside of Amazon
Full: aboutamazon.com/news/company…
Noticed some company linked to this EF Migrations article from their JIRA (via web analytics) and it annioyed me that it had errors so...UPDATED mostlylucid.net/blog/efmigra…
This is JUST me covering all the ways you can do migrations with EF and condenses all the best practices into one short piece.
