WEBVTT 00:00:00.000 --> 00:00:03.934 Two rival camps at the front of the AI race spoke of easing off in the same week. 00:00:03.934 --> 00:00:07.868 Is that real concern for safety? 00:00:07.868 --> 00:00:11.802 Or is it a new move in the competition? 00:00:11.802 --> 00:00:15.736 We start from that question. 00:00:15.736 --> 00:00:19.670 Then we follow one thread through regulation, leaked technology, and the flow of 00:00:19.670 --> 00:00:23.604 money. 00:00:23.604 --> 00:00:27.538 In each case, the words have moved ahead of the action. 00:00:27.538 --> 00:00:31.472 Concrete plans have not caught up. 00:00:31.472 --> 00:00:35.406 We ask what that gap means for people who develop, review, and promote medicines. 00:00:35.406 --> 00:00:39.340 It matters because the tools you plan to rely on are shaped by these choices. 00:00:39.340 --> 00:00:43.087 A remark about easing the pace surfaced this week. 00:00:43.087 --> 00:00:46.833 What matters is where it was said. 00:00:46.833 --> 00:00:50.580 It was said inside the company, to employees. 00:00:50.580 --> 00:00:54.326 A public statement is carefully worded. 00:00:54.326 --> 00:00:58.073 Words meant for staff tend to carry more of what leaders really think. 00:00:58.073 --> 00:01:01.819 So this may be more than a public pose.Caution came from several directions at 00:01:01.819 --> 00:01:05.566 almost the same time. 00:01:05.566 --> 00:01:09.312 One voice alone would be a personal view. 00:01:09.312 --> 00:01:13.059 When people in different positions lean the same way, the mood of the industry is 00:01:13.059 --> 00:01:16.805 shifting. 00:01:16.805 --> 00:01:20.552 Each remark came with conditions attached. 00:01:20.552 --> 00:01:24.298 Still, when the direction lines up, listeners change how they act. 00:01:24.298 --> 00:01:28.044 Rivals showed a similar stance within days. 00:01:28.044 --> 00:01:31.791 Coincidence or shared pressure, the signal to users is the same.For drug makers, the 00:01:31.791 --> 00:01:35.538 vendors behind the tools you are evaluating have turned cautious. 00:01:35.538 --> 00:01:39.284 When suppliers change their stance, your planning assumptions change too. 00:01:39.284 --> 00:01:43.031 A tool you meant to build into development could be put on hold by its maker. 00:01:43.031 --> 00:01:46.777 Having a fallback is what separates teams here. 00:01:46.777 --> 00:01:50.524 So why has this caution come up now? 00:01:50.524 --> 00:01:54.270 The answer lies in how several events piled up at once. 00:01:54.260 --> 00:01:57.927 The caution did not appear from nowhere. 00:01:57.927 --> 00:02:01.594 Several events landed in the same period, and how they overlapped is the point. 00:02:01.594 --> 00:02:05.261 Reports of abuse, people leaving, outside warnings, and new law. 00:02:05.261 --> 00:02:08.928 Each is a different kind of pressure. 00:02:08.928 --> 00:02:12.595 They moved almost together. 00:02:12.595 --> 00:02:16.262 Small on their own, together they shift the mood of a whole industry.Abuse, 00:02:16.262 --> 00:02:19.929 departures from inside, and legislation arrived at once. 00:02:19.929 --> 00:02:23.596 Any one of them can be handled. 00:02:23.596 --> 00:02:27.263 Together they create pressure to pause. 00:02:27.263 --> 00:02:30.930 When a person responsible for safety leaves a developer, it weighs more than outside 00:02:30.930 --> 00:02:34.597 criticism. 00:02:34.597 --> 00:02:38.264 The concern comes from inside. 00:02:38.264 --> 00:02:41.931 Once the reasons are public, those who stay find silence harder. 00:02:41.931 --> 00:02:45.598 Warnings from well-known outside researchers add to it. 00:02:45.598 --> 00:02:49.265 Pressure now comes from inside and outside. 00:02:49.265 --> 00:02:52.932 This pattern helps you predict what a vendor will do. 00:02:52.932 --> 00:02:56.599 Count the kinds of pressure, and you can guess the next move. 00:02:56.599 --> 00:03:00.266 The same dynamics can hit drug companies when technology, people and rules shift 00:03:00.266 --> 00:03:03.933 together. 00:03:03.933 --> 00:03:07.600 Next, we look inside the law and the leak. 00:03:07.610 --> 00:03:11.835 A way to steal technology, and a system to watch over it. 00:03:11.835 --> 00:03:16.061 Both were reported on the same day. 00:03:16.061 --> 00:03:20.286 One alone tells half the story. 00:03:20.286 --> 00:03:24.511 Side by side, they show that attackers and defenders both have gaps.An accusation, a 00:03:24.511 --> 00:03:28.736 new law, and a pointed objection arrived together. 00:03:28.736 --> 00:03:32.962 Now both sides are tested on substance. 00:03:32.962 --> 00:03:37.187 A defensive frame with nothing inside cannot stop an attack. 00:03:37.187 --> 00:03:41.412 Building the frame and making it work are separate jobs.One company's work was being 00:03:41.412 --> 00:03:45.637 used to skip the expense of building a model. 00:03:45.637 --> 00:03:49.863 If copying is cheap, the firm that paid for the original cannot earn that money 00:03:49.863 --> 00:03:54.088 back. 00:03:54.088 --> 00:03:58.313 The basis of competition breaks down.This matters for pharma too. 00:03:58.313 --> 00:04:02.538 Your research data could be drained through the same kind of route. 00:04:02.538 --> 00:04:06.764 Using AI and protecting what feeds it can no longer be planned separately.The count 00:04:06.764 --> 00:04:10.989 of accused groups points to a structural problem, not a single bad actor. 00:04:10.989 --> 00:04:15.214 One case is an exception. 00:04:15.214 --> 00:04:19.439 At this scale, the response must be structural too. 00:04:19.439 --> 00:04:23.665 The law meant to protect arrived as an empty frame. 00:04:23.665 --> 00:04:27.890 What you prepare for that gap leads straight into where the money is going. 00:04:27.890 --> 00:04:32.405 Money keeps moving while the safety debate goes on. 00:04:32.405 --> 00:04:36.919 The voices urging caution and the hands signing checks belong to the same industry. 00:04:36.919 --> 00:04:41.434 But the flow is uneven. 00:04:41.434 --> 00:04:45.949 Where it pools and where it never reaches tells you how the industry is built.There 00:04:45.949 --> 00:04:50.464 is big capital and a small grant, and nothing in between. 00:04:50.464 --> 00:04:54.978 That gap works as a barrier to entry. 00:04:54.978 --> 00:04:59.493 Mid-sized firms are stuck. 00:04:59.493 --> 00:05:04.008 They cannot raise huge sums alone, and they rarely qualify for public grants. 00:05:04.008 --> 00:05:08.522 That squeezed middle is probably the largest group.The smaller recipient is building 00:05:08.522 --> 00:05:13.037 systems that work without step-by-step human direction. 00:05:13.037 --> 00:05:17.552 Whether that lets a lean team compete is still unknown. 00:05:17.552 --> 00:05:22.066 If there is a path to results on a small budget, this is where to look. 00:05:22.066 --> 00:05:26.581 Proof will take time.The larger raise is expected to go into equipment and people. 00:05:26.581 --> 00:05:31.096 The size of the money is becoming the wall itself. 00:05:31.096 --> 00:05:35.611 Capital, more than technical skill, now decides who can compete. 00:05:35.611 --> 00:05:40.125 This split in funding ties directly to the surge in users we turn to next. 00:05:40.125 --> 00:05:44.640 More users means bigger infrastructure bills. 00:05:44.640 --> 00:05:49.133 The builders talk about caution, yet user demand does not stop. 00:05:49.133 --> 00:05:53.627 This week the contradiction took a concrete form. 00:05:53.627 --> 00:05:58.120 Users rushed in, and registration itself was halted. 00:05:58.120 --> 00:06:02.613 The company that spoke of braking could not keep up with its own customers.Requests 00:06:02.613 --> 00:06:07.107 outran capacity, and the door was closed for a while. 00:06:07.107 --> 00:06:11.600 Turning away willing customers was likely a surprise for the provider. 00:06:11.600 --> 00:06:16.093 It costs revenue, and it can cost trust. 00:06:16.093 --> 00:06:20.587 Meanwhile, experts objected to how a scientific claim was handled. 00:06:20.587 --> 00:06:25.080 Speed of growth and checking of results are out of step. 00:06:25.080 --> 00:06:29.573 Building fast and building right are different problems.The experts objected to the 00:06:29.573 --> 00:06:34.067 process, not the result. 00:06:34.067 --> 00:06:38.560 Publishing without proper checks damages trust in the technology. 00:06:38.560 --> 00:06:43.053 In drug development, it is like claiming efficacy while skipping human trials. 00:06:43.053 --> 00:06:47.547 Skipping steps is faster. 00:06:47.547 --> 00:06:52.040 But lose trust, and no one will use what you built. 00:06:52.040 --> 00:06:56.533 The short-term gain comes with long-term damage to credibility. 00:06:56.533 --> 00:07:01.027 Keeping both demand and trust is now the developer's test. 00:07:01.027 --> 00:07:05.520 That question of trust reaches straight into pharma work. 00:07:05.520 --> 00:07:09.179 How do these moves touch the daily work of making medicines? 00:07:09.179 --> 00:07:12.838 Adoption, control and data protection each carry problems. 00:07:12.838 --> 00:07:16.497 Fix one and the others stay open. 00:07:16.497 --> 00:07:20.156 The caution talk, the leak accusation and the funding gap look separate. 00:07:20.156 --> 00:07:23.814 Their roots connect. 00:07:23.814 --> 00:07:27.473 They all end at the decisions your teams make every day.The center of gravity is 00:07:27.473 --> 00:07:31.132 moving. 00:07:31.132 --> 00:07:34.791 The question used to be whether to bring tools in. 00:07:34.791 --> 00:07:38.450 Now it is how to manage them once they are in. 00:07:38.450 --> 00:07:42.109 Organizations with room to manage are pulling ahead. 00:07:42.109 --> 00:07:45.768 The copying method we saw earlier could also reach your internal research data. 00:07:45.768 --> 00:07:49.427 So adoption, control and data protection need one design. 00:07:49.427 --> 00:07:53.086 Any team can use a tool. 00:07:53.086 --> 00:07:56.744 Taking responsibility for what happens next needs people and budget. 00:07:56.744 --> 00:08:00.403 Adopt without a management frame, and you will react only after an incident. 00:08:00.403 --> 00:08:04.062 Build the frame while you still have room. 00:08:04.062 --> 00:08:07.721 Uneven adoption and missing control now overlap. 00:08:07.721 --> 00:08:11.380 That makes it worth checking what the news leaves unsaid. 00:08:11.380 --> 00:08:14.701 Several conditions sit behind today's events, out of view. 00:08:14.701 --> 00:08:18.022 They matter for your judgment. 00:08:18.022 --> 00:08:21.343 News reports cover what moved. 00:08:21.343 --> 00:08:24.664 They often skip the caveats. 00:08:24.664 --> 00:08:27.985 Every event today is still at the announcement stage. 00:08:27.985 --> 00:08:31.306 What was intended and what will actually happen may differ. 00:08:31.306 --> 00:08:34.627 So you need to sort what is backed by evidence from what is still up in the air.What 00:08:34.627 --> 00:08:37.948 they share is missing yardsticks and missing outside confirmation. 00:08:37.948 --> 00:08:41.269 There is distance between words and action. 00:08:41.269 --> 00:08:44.590 The side making the charge competes with the accused. 00:08:44.590 --> 00:08:47.910 The law has no content yet. 00:08:47.910 --> 00:08:51.231 Expert judgment is not settled. 00:08:51.231 --> 00:08:54.552 It is too early to conclude anything. 00:08:54.552 --> 00:08:57.873 If you judge by announcements alone, you will drift from reality. 00:08:57.873 --> 00:09:01.194 Waiting is not doing nothing. 00:09:01.194 --> 00:09:04.515 It means deciding in advance what would make you act. 00:09:04.515 --> 00:09:07.836 Then the next headline will not push you around. 00:09:07.836 --> 00:09:11.157 Write those triggers down. 00:09:11.157 --> 00:09:14.478 That is a practical step you can take today. 00:09:14.478 --> 00:09:17.799 Mood is moving ahead of the yardsticks. 00:09:17.799 --> 00:09:21.120 Separating what is settled from what is not will guide the decisions still to come. 00:09:21.120 --> 00:09:24.806 Across today, one pattern held. 00:09:24.806 --> 00:09:28.491 Words moved first, and substance trailed behind. 00:09:28.491 --> 00:09:32.177 Leaders promised care without a plan. 00:09:32.177 --> 00:09:35.863 A law passed without standards. 00:09:35.863 --> 00:09:39.549 A charge was made without independent proof. 00:09:39.549 --> 00:09:43.234 Each event left a gap between stated intent and action. 00:09:43.234 --> 00:09:46.920 Whether those gaps close is what we will watch in the coming days.