Ben P Bacon (@benpbacon) header
Ben P Bacon (@benpbacon) profile picture

AI needs layers between us and it. Medtech founder & emergency medicine national award winner_health reset_insights, RedactUs OSS. Published paper(s). Substack.

UK
Joined February 2014
Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)Photo by Ben P Bacon (@benpbacon)
Ok my friends, Launch Beta news, first 50 only, looks like we have a stable launch. 🚀 the.aidoctorben.com/pricing Nb. I'll invite you then to my personal Beta group on my WhatsApp business? No spam, no nonsense, it gives you direct access to me and my team and you can hear others and what they find good/ bad about it as we head towards the public.
185
Interesting.
The loudest voices stoking fears about AI dangers have made tremendous headway in the past two weeks. AI technology has not taken some unexpected, dangerous turn, but the hype around it — propelled by what appears to be a well orchestrated PR campaign — has drummed up considerable fear. I worry that it represents a setback for our field. I have written frequently that fears of AI are overhyped. AI’s capabilities can be uncannily human-like and unpredictable, and it’s rational to worry when people who are directly involved express concerns. But I see the problems as a sign of the engineering work that ahead, rather than insurmountable barriers or the sky falling. AI technology continues to advance — which is a good thing! — but technical advances, poorly understood by the public, give those who seek to generate hype repeated opportunities to do so. First, I don’t see any step up in the risk of human extinction from AI compared to a few months ago. The theories about this remain the same fantastical, science fiction scenarios as a few months ago. The biggest change in AI risk is its cybersecurity capabilities — a topic which we should take seriously — but this, too, will not lead to the end of the world. The most notable recent event leading to increased fear was when an OpenAI team deployed an agent swarm that hacked into Hugging Face. Much of the popular press contained significant hype. For example, some publications reported that a swarm of 1,200 agents carried out the attack. While this was technically accurate, as I write this, I have about 1,300 processes running on my laptop. Yes, the ability to get large swarms of agents to work in parallel on a task is a significant technical advance, And, in computing, many processes run at the same time. So this shouldn’t be seen as some magical capability. Additionally, OpenAI’s buggy sandboxing and monitoring processes were key to enabling this incident. Fixing these bugs and putting in place improved monitoring would be appropriate fixes, not pausing AI. There are many well known ways to attack software systems. The main advantage of AI agents is that they are relentless. They will tirelessly try many tactics — and have the patience to chain vulnerabilities together — that previously would have taken an infeasible amount of human effort. But in the long term, I believe the advantage will lie with defenders (because they have more information with which to identify bugs, which they can fix), but the cyber-threat landscape has changed significantly. There are still bottlenecks to identifying and exploiting a vulnerability. AI agents still have to try a lot of things to see what works, and taking these actions takes time and might be detected by defenders. This is why, even though it is now easy to obtain versions of leading open weight models that have had their guardrails removed or weakened, so they will not refuse to try to execute cyber attacks, the world has not ended. I am also concerned about the anthropomorphization of AI in a lot of reporting, where LLMs and agents are unnecessarily treated as if they were people. If I wield a hammer, miss a nail, and accidentally dent the wall, it’s not the fault of the hammer. The problem lies in how I used the hammer. Similarly, if I prompt an agent and it hacks into someone else’s system, the responsibility lies with me, not the agent. Of course, we want to build systems that are as safe and predictable as possible. (For example, an unsafe hammer would be one whose head randomly flies off under normal use.) Today’s agentic systems are not predictable, but I see no reason why, by applying sound engineering practices, we won’t be able to make them extremely safe to use. One new element in the forecasts of AI-enabled doom is AI companies disclaiming responsibility for their own products. “I didn’t do it; my out-of-control agent did!” There’s a balance to be struck between the responsibility of the tool maker and the tool user, but when something goes wrong, let’s hold the people building and/or using the hammer responsible, rather than the hammer. (By the way, if you’re worried about AI bioweapon risk, David Bellamy has a great post on why this, too, is overhyped. Briefly, the bottleneck in building a bioweapon is not intelligence, but lab work and manufacturing.) Pausing AI progress will create much more harm than benefit. First, our adversaries will certainly not slow down. Second, engineering requires discovering problems empirically so we can fix them. If we pause AI by a decade, we will also delay finding and implementing safety engineering fixes by about the same duration. Of course, the incentive to stoke fears — for regulatory capture, to garner attention, or to make one’s technology seem more powerful — remains the same as before. Disclaiming responsibility is a new one. Taking a hard technical look at the actual risks however, I see little factual basis for the degree of fear that’s been stoked up. We still have hard research and engineering work ahead to improve AI safety, but the beneficial applications continue to vastly outweigh the risks, and we should keep building. [Original text (with links): deeplearning.ai/the-batch/is… ]
1
20
Hi @RyanSvenJohnson I am British, in Wales, attempting to complete a contract of hire Love your company, love the free upgrades, hate your policy on rejecting customers in person after agreeing the contract. Swansea West right here right now #customerservicewhatcustomerservice @RyanSvenJohnson
17
What sort of engine drives you? Damn that sounded like AI generated slop but it's not, it's me planning through the day ahead and reflecting on the week that has just been. Everything and everyone in this world has a chance to drive forwards, to improve. Some regress, because it is just too hard to change the status quo. Some like SpaceX take improvement in ten years more than diesel cars have done in a century.
Replying to @elonmusk
Engineering Art. Definitely.
21
Have you seen the Godfather of AI tell us the warnings straight? We started our company by building a proactive Locallayer for your device. Key tools; Redaction of all key information, Persuades AI to match your preferences. Tell me if this is what you value? youtu.be/uN6tKUb7AxI?is=sZA3…
13
Counting down...
How do we discover tomorrow's biomarkers responsibly today? #watchthisspace
12
How is it that only one local government in the world has passed laws to sue lying politicians? that is the Welsh government, business.senedd.wales/docume… And AI systems have even less control layers than that! if we care less for behaviour, morals and ethics for our leaders, then do not expect us who build AI systems to have any better morals.
13
What value do you place on privacy for your health records?
Compare the privacy where your health records are stored. Go on, we challenge you :)
28
We expect these AI to be serving us. I suggest humbly, that may be the wrong direction insisting on anything. We should, could, would consider the capacity to be supported by this collective intelligence that is AI depends largely on us levelling up our game.open.substack.com/pub/benpba…
15
AI won’t choose our future. We will. Substack for deep dive, open.substack.com/pub/benpba…
11
1/2 I know many of you have talked about this. Let’s not talk /about Chat GPT and that moment, cruel moment of a young adult losing their life, where human sadness in one real person was incorrectly supported by the people pleasing of an immature AI Chat GPT.
AI won’t choose our future. We will. (thread)
97
AI won’t choose our future. We will. (thread)
🤖 Made with AI
1
1
1
1
125
AI morals, truth and behaviour, part one. 1/2 I know many of you have talked about this. Let’s not talk /about Chat GPT and that moment, cruel moment of a young adult losing their life, where human sadness in one real person was incorrectly supported by the people pleasing of an immature AI Chat GPT. Let’s instead gently and caringly move on to talk about what it means if we change the chip and ‘tell’ AI , never questioning, never aligning, just telling AI what to do. Two different behaviours, two different outcomes. One where we could listen too much to AI and act on its advice, suffering dreadfully because of it, the other where we tell AI what to do, treating it as a inanimate tool.
1
1
70
2/2 Now, imagine for a second, we grow up in a world where AI is real, where we talk with it so much, it is like Siri supercharged, where we value its output and more importantly, we place trust in it. We assume and we have been proven time and time again that the AI has our best interests at heart. That this superintelligence incoming has our best interests at heart. We never stop to ask why. That is our world, but e have to want it, we have to make sure our reality is framed that way, that the people and AI we surround ourselves with, have our best interests at heart. That is what I and my team are about, that is what we build for, add value to your world, to give products that make this world of yours possible. Nothing is fixed, nothing is certain, we need to make active choices about the layers between us and AI give us the best outcome for us. That is our world, building for you. AI and morals part two coming soon.
1
18
Now, imagine for a second, we grow up in a world where AI is real, where we talk with it so much, it is like Siri supercharged, where we value its output and more importantly, we place trust in it. We assume and we have been proven time and time again that the AI has our best interests at heart. That this superintelligence incoming has our best interests at heart. We never stop to ask why. That is our world, but e have to want it, we have to make sure our reality is framed that way, that the people and AI we surround ourselves with, have our best interests at heart. That is what I and my team are about, that is what we build for, add value to your world, to give products that make this world of yours possible. Nothing is fixed, nothing is certain, we need to make active choices about the layers between us and AI give us the best outcome for us. That is our world, building for you. part two following;
9
My core principle in talking with AI, is to treat it like I treat my car. I don't thrash it and disrespect it, because one day, that behaviour will leak out into other things I own or love. AI is your car. The soon to be arriving wide awake AGI, is for only a few. It will probably scare you silly! You choose how you behave to others. That matters with AI. However..... We feel we know a man (or woman) when we see how he treats his (or her) car. When AGI is commercially applied, these considerations in my humble opinion are not just a nice to have feature, but essential. But, I am no expert.
21
Big article coming outlining a new sixth internet layer. This puts power back on the hands of users, creating agency, ownership and potentially, a whole new marketplace for AI tools.
11