https://www.theregister.com/2026/02/18/palo_alto_q2_26/
If enterprises are implementing AI, they're not showing it to Palo Alto Networks CEO Nikesh Arora, who on Tuesday said business adoption of the tech lags consumer take-up by at least a couple of years – except for coding assistants.
"Consumers are far outstripping enterprise for the moment, but we expect enterprise will surely and slowly get on that bandwagon," he said on the company's Q2 earnings call.
Arora likened business uptake of AI to the cloud computing shift, which he said took two or three years before enterprises started migrating applications.
"Right now ... tell me how many enterprise AI apps are you using which are driving tremendous amounts of throughput," he asked, and answered himself "I can't think of anything but coding apps."
Coding apps aren't great for Palo Alto's business because they don't generate a lot of network traffic to which it can apply its security smarts. Arora thinks his security vendor peers know this.
"We're all laying the groundwork right now. It is ... sort of an arms race to try and see who can get the AI security sort of platform up and running as quickly as we can."
But the limited enterprise AI adoption Arora has seen does pose some immediate challenges to Palo Alto.
"There is now enterprise adoption that we're beginning to see where customers are running perhaps millions of tokens in one or two particular applications they're working with some of the LLM providers on, and that's where we see the traffic," he said. That traffic is on the LAN and the CEO doesn't think existing networks struggle to handle it.
"I think the challenge right now is consolidating that traffic," he said. "How do you get all the AI traffic to be in one place? So you can understand it, provide visibility, look at the ability to control it and be able to act on it."
The CEO said that as this sort of AI-related traffic grows "it needs a different set of controls and tools."
Palo Alto is already getting its hands on those tools, as on Tuesday put to bed rumours it would acquire agentic AI endpoint security startup Koi by announcing it's done the deal.
Arora pointed to Palo Alto's recent acquisitions of Chronosphere and CyberArk as further evidence of the company's moves to ensure it builds a portfolio of products to secure the AI enterprises will eventually implement.
The CEO expressed confidence Palo Alto has the products it needs today, saying customers know they can't prepare for AI if they are running a tangle of security tools and are therefore consolidating to the kind of platforms the company offers.
Demand for those products helped Palo Alto to $2.6 billion Q2 revenue for the quarter, which represented 15 percent year-over-year growth.
Execs pointed to the success of the company's subscription offerings, noting 23 percent growth in remaining performance obligations, which now stand at $16 billion. And they predicted Q3 revenue would grow at least 28 percent to land between $2.941 billion and $2.945 billion.
All of those nice numbers didn't impress investors, who knocked six percent off the company's share price – perhaps because they weren't thrilled by predictions that profits will ease.
(Score: 0) by Anonymous Coward on Saturday February 21, @05:21AM
The computer industry keeps churning so that someone gets to try to
grab the brass ring, like Google did with monetizing search.
Fuck them all...
(Score: 3, Interesting) by looorg on Saturday February 21, @02:45PM (5 children)
I'm sure it will be great for business. Soon (TM). Greatness is about 10 years/months away. Always as a constant.
It does seem a bit vague tho regarding what it is that they are securing. Are they securing the AI from outside interference? Or from its own idiocy or from the users? Or are they making sure the output is inline with the right company values etc ... Perhaps the answer to that is just YES, All of the above and more.
He forgot the most obvious one. So that we can monetize it. Preferably also they should put all the AI in one place so that it can be nuked. It's easier to block it for the rest of us if it's all confined to one place. More of that.
They can't detect any patterns in the current level of idiocy? One would think it would be evident by now.
(Score: 0) by Anonymous Coward on Saturday February 21, @02:49PM
Are they hardening their new Forbin-inspired info bunkers from a localized EMP?
just... uh, curious.
(Score: 4, Interesting) by JoeMerchant on Saturday February 21, @04:41PM (2 children)
This guy is a little on the enthusiastic side, but also not far wrong in my opinion: https://www.youtube.com/watch?v=bDcgHzCBgmQ [youtube.com]
The thing that people miss about "job skills" is that most of what goes into job skills isn't what you learn in academia, or your skills in a particular craft or trade, most of what gets you hired and paid at US labor market rates is: communicating with US customers. In any size company, but especially large ones, your co-worrkers (especially your management) are your customers. When you hire a general contractor, he hires sub contractors, and they do the labor - they charge for materials plus their time, but most of the time I see home construction / remodling sub contractors bidding most jobs at 2x materials cost - it's like the 15 (now 20)% tip - they know the market will bear that cost because they're spec'ing the material and obviously they don't want to do the labor themselves... anyway, the GC then takes the sum of his sub-bids, doubles that (or more, if the customer, aka mark, looks like they will pay), and that's their bid. What's the GC's value in this relationship - mostly: communication (aka management). In decades gone by, I thought a management skill was the ability to step in and perform the work of the managed when the managed don't show up or aren't performing at required levels. I see that less and less these days - now management mostly seems to mean: communication between those who do and those who pay. It's a pyramid scheme: you be my point of contact with the rabble, and I'll pay you so I don't have to communicate with all of them - just you.
Even when you're rabble, your primary value to your management is that they can communicate with you, as efficiently as possible, and also that you do (or get done) what they're asking for, which is what their management is asking for, which is what their management is asking for. It may be turtles all the way down, but it's turkeys all the way up.
If your management can communicate directly with AI and actually get it to do what they've been asking you to do all these years - and do that as or more efficiently as they communicate with you - you're hosed. Just like the typing pool went away in the 1980s. Just like the transcriptionists went away in the late 1990s. Just like horse drawn carriages went away in the 1920s.
Right now, I don't see my management being able to drive AI to do my work for less effort than they drive me to do the work. Yeah, they can get AI to do what I do, but it "works them" a lot harder than I do. That's where your value as a white collar employee lies (and always has) - like domestic staff: you just make it happen.
🌻🌻🌻🌻 [google.com]
(Score: 2) by VLM on Saturday February 21, @05:14PM (1 child)
I'd agree with about 99% of that, just to start with.
WRT the end of your post I think you're describing being a subject matter expert. I had a client who wanted some programming and automation done for a "Heavy RF" project in the engineering dept at a heavy RF company and it helped that I knew exactly what they are doing, why they were doing it, and how to do it, before I met with mgmt. They don't need to prompt me with what a dBmW unit is, or why you'd monitor the 3rd order intercept point of an amplifier, or how you'd use a recent model keysight spectrum analyzer (which they already owned) to do it. At a high enough level the CEO probably has no idea what the word definitions mean. At a slightly lower level a manager wouldn't know what concepts and language to use. I was working on revision 2 with the QA lab technicians before an upper level manager could have (probably incorrectly) completed a creative writing prompt. Everything turned out really well with this project for everyone, they even paid my invoice pretty quickly. The gain from a human subject matter expert over AI is the time NOT used having to write a prompt that management probably doesn't understand enough to write anyway.
Earlier in your response I'd consider removal of headache as a GC feature. The GC is being paid to remove headaches from whomever hired him. If there were 36 hours in a day the GC wouldn't have a job, but some dude has 16 hours of headaches per day and a GC is willing to get rid of 8+ hours worth for a fee they can afford so... Project management as a service is a legit human job. My experience with clients is they generally aren't interested and I'm better off with them not knowing if a specific project was a lot of effort or not. They just want the pain to go away and I agreed to do it for $10K or whatever and we both think we got the better side of the deal. Sometimes you'll hear managers tell direct reports stuff like "they don't want to hear problems they want to hear solutions" but my experience in non-W2 contract work is they don't want to hear anything about nothing, here's a pile of $$$ now make this headache disappear. I have a great-uncle who got into fine interior woodworking like custom cabinets and he said something about his line of work wasn't really being paid to build stuff like outsiders would think, but being paid to demolish problems, he wasn't really paid to slap cabinets together at the jobsite he was getting his $$$ for making cabinet related problems disappear from the jobsite, which admittedly often involved slapping together a lot of cabinets. But he really made his money by fixing things without bothering the owners or often even the GC, like if the plumbing rough in does not match the plans, theres the technically correct way to fix it, the baling wire and duct tape way to fix it, the "terrorize upper management" way to fix it, and his weird ways of knowing a guy who knows a guy who knows just what to say to make it all work and nobody hears about it, other than he got a fat paycheck for being really good at that compared to other carpenters.
(Score: 2) by JoeMerchant on Saturday February 21, @05:38PM
>Project management as a service is a legit human job.
Absolutely. Guy we bought this house from was a GC, he owned and lived in the house next door. As a GC, he was used to using 'his helpers." Our gate fell off (while we were renting for a few months before purchase closing). He tossed his helpers a carton of cigs and told 'em to go drive 35 miles (each way), pick up an new gate, and install it. Which they did, poorly, but good enough for him to get his renters (us) off his back and all it took was a "go do this" prompt and he was clear of the problem.
Presently, I'm abusing Claude to do some CAD work, drawing a crazy floorplan for a 3D printed concrete mini-home. I came to the conclusion that I wanted to rotate the coordinate system by atan(1/9) - yeah, I could have learned (and probably had to purchase for about what I'm paying Claude - possibly a lot more) some better CAD software where that was a click and drag maneuver - that would have taken me 5 to 50 minutes to figure out the first time, but instead I gave Claude a single prompt and it's chewing through the problem for me, looks like it's going to be about 20-30 minutes. It's managing all the detail for me - I'll see if I can copy-paste some of it's internal monologue at the end of this post... Point being, I'm screwing around reading your stuff and writing mine (relatively fun) while Claude is crunching through the problems, untangling the dependencies, etc.
>they generally aren't interested and I'm better off with them not knowing if a specific project was a lot of effort or not. They just want the pain to go away and I agreed to do it for $10K or whatever
Starting with my first "consulting" gig in college I began to realize: some of the simplest things are the most valuable. I got paid $50 to show up and tell some guy that the address of the main drive in DOS was C: - I felt kinda guilty, but he was incapable of explaining his issue (he just kept repeating that he couldn't get the backup software to run) over the phone and I did drive out to meet him at his place on his schedule, so...
> fixing things without bothering the owners or often even the GC
Yep... here's the snippet:
Another error! This time at gen_floorplan.py:971: wh_n = pts["C7"][1] + math.sqrt(wh_tangent_r**2 - wh_dE**2). The issue is wh_tangent_r**2 - wh_dE**2 0, meaning the water heater circle doesn't intersect the C7 arc.
Let me understand this. The water heater is positioned tangent to the C7 arc (which is the arc at the W8-W9 corner). After rotation, the positions have changed, and the geometric relationship might have changed enough that the tangency computation fails.
I'm tracing through the tangency calculation: the water heater center's E-coordinate is derived from the inner wall position plus its radius, the tangent distance accounts for the arc radius minus wall thickness minus the water heater radius, and the E-distance measures how far the center sits from the arc's center point. The N-position then uses the Pythagorean relationship to find where the circle touches the arc, but the math breaks down when the center is too far from the arc to maintain tangency.
The real problem is likely a mismatch in coordinate systems — some values like the IW2 easting come from non-rotated constants while the arc points themselves get rotated, so their relative positions shift enough to violate the tangency constraint. In the pre-rotation system, IW2's east face is around E ≈ 7 while C7 sits at E ≈ 10, making the distance wh_dE ≈ -1.83 (IW2 west of C7). But when I rotate the coordinate system by ~6.34°, the E-coordinates shift differently for each point, potentially flipping the sign or magnitude of that distance.
Another issue in gen_floorplan.py — water heater placement fails after rotation. Let me check the actual values.
Bash Check water heater positioning values
IN
cd /c/Users/Mango\ Cat/Dev/hut2 && python -c "
from floorplan.gen_floorplan import build_floorplan_data
from floorplan.constants import WH_RADIUS
import math
data = build_floorplan_data()
pts = data.pts
layout = data.layout
wh_e = layout.iw2.e + WH_RADIUS
wh_tangent_r = (data.radii['R_a7'] - data.wall_t) - WH_RADIUS
wh_dE = wh_e - pts['C7'][0]
print(f'iw2.e = {layout.iw2.e}')
print(f'WH_RADIUS = {WH_RADIUS}')
print(f'wh_e = {wh_e}')
print(f'C7 = {pts[\"C7\"]}')
print(f'wh_dE = {wh_dE}')
print(f'wh_tangent_r = {wh_tangent_r}')
print(f'wh_tangent_r^2 - wh_dE^2 = {wh_tangent_r**2 - wh_dE**2}')
"
OUT
iw2.e = 6.83435271088773
WH_RADIUS = 1.1666666666666667
wh_e = 8.001019377554396
C7 = (5.33752376028426, 23.729469739518176)
wh_dE = 2.6634956172701365
wh_tangent_r = 0.5000000000000002
wh_tangent_r^2 - wh_dE^2 = -6.8442089032172255
The issue is clear: wh_dE = 2.66 but wh_tangent_r = 0.5. So the water heater center is 2.66' east of C7, but the tangent radius is only 0.5'. The water heater is way too far east of C7 for tangency.
Let me check what C7 was in the pre-rotation system:
Pre-rotation C7: In _EXPECTED_F, F8 = (10.2587, 22.9949). C7 is the arc center near F7-F8-F9. Looking at the outline, C7 is at roughly E=10, N=23.
Post-rotation C7: (5.34, 23.73). So C7 moved from E≈10 to E≈5.34 after rotation.
Now I'm tracing back where the inner wall 2 position comes from. W2 is near F2 on the west side, which was at E≈-0.74 pre-rotation and E≈-0.83 post-rotation. The inner wall 2 east coordinate is calculated by adding the offset and wall thickness to W2's position, giving me roughly E≈-0.07 pre-rotation.
Looking at the water heater position relative to C7, I can see the relationship flipped between the two coordinate systems. Pre-rotation, the water heater was west of C7 with a negative offset of about -2.16, but post-rotation it's east of C7 with a positive offset of about 2.66.
The root cause is that C7 sits in the upper part of the building near F8, while the water heater's E-coordinate comes from IW2 in the lower section near F2. When rotating counterclockwise around F1 at the origin, points higher up shift much more eastward than lower points do. So C7 moved west significantly (from 10.26 to 5.34) while the water heater barely shifted (6.93 to 6.83), completely breaking the geometric assumptions the placement algorithm relies on. longer hold after rotation.
Actually, I think the issue might be with how IW2 is computed. Since IW2 is positioned based on W2's easting plus an offset, and W2 shifts when the building rotates, the real problem is the relative positioning of C7 versus IW2 in the east direction. Before rotation, the west wall was nearly vertical, so IW2 sat east of it while C7 was well further east. But after the ~6.34° counterclockwise rotation, C7 at the top of the building shifts significantly westward—dropping from E=10.26 to E=5.34—while IW2 near the bottom of the west wall barely moves. This is a genuine geometric consequence of the rotation, so I need to rethink how the water heater gets positioned.
Instead of anchoring the water heater to a fixed E-coordinate relative to IW2, I should position it tangent to the C7 arc in a way that adapts to the rotation. Calculating tangency... rotation, the E-coordinate of C7 has changed relative to IW2.
A better approach: compute the water heater center position based on IW2 and the north wall, rather than tangency with C7. Or, compute it from the rotated geometry directly.
Actually, looking at this more carefully, the water heater is placed tangent to the inner W8-W9 arc (C7 center). It should be tangent from the IW2 side. In the pre-rotation system, the water heater was west of C7 (between IW2 and C7). After rotation, C7 is now WEST of the water heater center, which means the water heater can't be tangent from the west side.
The fix: compute the water heater position using the actual geometry of the rotated building. The water heater should be east of IW2 with its west face against IW2's east face, and tangent to the inner C7 arc. After rotation, C7 has moved relative to IW2, so the tangency computation needs to work with the actual positions rather than assumed coordinate relationships.
Wait, this is actually a rendering issue in gen_floorplan.py, not a geometry computation problem. The layout computation doesn't position the water heater — that happens in the rendering code. I'm looking at how the rendering calculates the water heater position using the IW2 east coordinate and the radius values.
The current approach positions the water heater at a specific coordinate where it's tangent to the inner C7 arc, but this only works if the distance from the water heater center to C7 is less than the tangent radius. After rotation, that distance is 2.66 but the tangent radius is only 0.5, so the water heater center is way outside the tangent circle.
The root cause: the water heater positioning uses IW2's easting plus the radius, but IW2's easting is no longer meaningful relative to C7 after rotation. I need to recompute the water heater position based on the rotated geometry.
Looking at the rotated coordinates, the water heater should sit between IW2's east face and be tangent to C7's inner arc. The inner arc radius is 1.67 and the water heater radius is 1.17, so the tangent distance should be 0.50. Let me recalculate where the water heater center actually needs to be positioned. The real problem is that in the rotated layout, IW2 ended up east of C7, so pushing the water heater further east only makes it worse. I need to compute the tangent position directly from C7 toward IW2 instead of just offsetting from IW2's edge.
Computing tangent position... I'm realizing the water heater positioning is really a rendering issue rather than a core geometry problem—the bug stems from how the code calculates the E-coordinate after rotation. Let me trace through the actual implementation to see where this is being used and find the cleanest fix.
🌻🌻🌻🌻 [google.com]
(Score: 4, Interesting) by VLM on Saturday February 21, @04:47PM
This is one of those articles that was dumbed down by the journalism filter and/or the AI filter making it hard to understand than if it was not dumbed down.
However, given the above, I "think" that the article is talking about companies that do dumb things like install a MCP interface on an internal server then expose it to the entire internet or too much of the internet to do AI slop stuff.
If you're not already involved, MCP is analogous to a REST API for AI LLMs to connect to. So rather than cut and paste your entire database schema attached to every prompt you connect an LLM once to a database server using an MCP interface. In theory the people doing this will be intelligent and security conscious, in practice we know how this turns out every time for the past couple decades. So instead of permitting read only access to one low security database's table schemas to one IP address, they'll give full read-write access to all databases including payroll and CC number storage and open it to the entire internet "because it worked, and it was too much work to secure".
And yeah unless you're doing something really expensive or really weird or both, MCP traffic is pretty light and intermittent.
The kind of AI slop you can do with a MCP server, on a DB anyway, would be "hey AI give me the exact SQL statement to join our illegal, forbidden by PCI-DSS table of unencrypted customer credit card numbers to our unsecured and unaudited list of customer orders that haven't been marked as billing process initiated" or something like that.
The kind of people who need AI slop to do basic coding are going to correlate strongly with the kind of people who can't secure their way out of a wet paper bag with a flashlight and map or something like that. Expect an absolute flood of "cybersecurity" stories relating to MCP, its going to be very fertile ground assuming the bubble lasts long enough.