With the Gemini 2.5 Flash release, Google envelopes the entire cost-quality frontier of LLMs. In other words, at any cost or quality level, today, the best model to use according to the LM Arena score is a Gemini model. Results for O3, O4 Mini, and GPT 4.1 are not yet on LM Arena. But until then, #Google dominates. Nice work! Link: https://sanand0.github.io/llmpricing/ LinkedIn
Looks like XML tags are the best way to structure prompts and separate sections for an #LLM. It’s the only format that all of Anthropic, Google, and OpenAI LLMs encourage. For example: … … … … Anthropic Docs: https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/use-xml-tags OpenAI Docs: https://platform.openai.com/docs/guides/prompt-engineering/strategy-write-clear-instructions Google Docs: https://cloud.google.com/vertex-ai/generative-ai/docs/learn/prompts/structure-prompts Alternatives are using JSON, Markdown, templating formats like Mustache/Jinja, etc. Even Llama’s system tokens seem a little XML-like. https://github.com/meta-llama/llama3/blob/main/llama/tokenizer.py#L61-L74 Personally, I’ve been using Markdown so far. But it’s time to switch over. (Only on the prompt side. On the generation side, Markdown still seems the best.) ...
I’m quite busy on a project right now, and don’t get time to write long articles. So for a while, I’m going to stick to short notes on interesting stuff. Peter Bregman has a very interesting piece on Why You Should Encourage Weakness. It boils down to a choice: do you focus on on improving strengths or minimising weaknesses? Conventional performance evaluations focus on the latter. I very strongly support Bregman’s view on this. The weakness isn’t why you hired the person! Unless it’s killing the organisation, just leave them to focus on their strengths. Google Analytics has a fairly interesting API that I hadn’t explored until recently. Picked up [Advanced Web Metrics with Google Analytics](http://www.s-anand.net/amazon-browser.html#advanced web metrics with google analytics) and learnt that you can track outbound clicks, page load times, Javascript events and error logs, almost anything at all using Google Analytics. You can also mirror the logging on your local server using pageTracker._setLocalRemoteServerMode() The whole concept of a Sandbox environment seems to be picking up within Google. There’s a Checkout sandbox, an AJAX API playground, an AdWords sandbox, an AdSense API sandbox, the Mapstraction API sandbox, even an event called Developer Sandbox. (After saying Sandbox 6 times, I feel a bit like Hobbes.)
One out of every 5 hits to my site is from a bot. I spent a fair bit of time this weekend analysing my log file for last month (which runs to gigabytes, and I ended up learning a few things about file system optimisation, but more on that later). 80% of the hits were from regular browsers. 20% were from robots. Here's a sample of the user-agents: Mozilla/5.0 (compatible; Yahoo! Slurp; <a href="http://help.yahoo.com/help/us/ysearch/slurp)">http://help.yahoo.com/help/us/ysearch/slurp)</a> Mozilla/5.0 (compatible; Googlebot/2.1; +<a href="http://www.google.com/bot.html)">http://www.google.com/bot.html)</a> Mediapartners-Google DotBot/1.0.1 (<a href="http://www.dotnetdotcom.org/#info">http://www.dotnetdotcom.org/#info</a>, [email protected]) Mozilla/5.0 (Twiceler-0.9 <a href="http://www.cuill.com/twiceler/robot.html)">http://www.cuill.com/twiceler/robot.html)</a> msnbot/1.1 (+<a href="http://search.msn.com/msnbot.htm)">http://search.msn.com/msnbot.htm)</a> FeedBurner/1.0 (<a href="http://www.FeedBurner.com)">http://www.FeedBurner.com)</a> Mozilla/5.0 (compatible; attributor/1.13.2 +<a href="http://www.attributor.com)">http://www.attributor.com)</a> WebAlta Crawler/2.0 (<a href="http://www.webalta.net/ru/about_webmaster.html)">http://www.webalta.net/ru/about_webmaster.html)</a> (Windows; U; Windows NT 5.1; ru-RU) Yandex/1.01.001 (compatible; Win16; I) ... You get the idea. The bulk of these are search engines. Over two-thirds of the bot requests were from Yahoo Slurp. Now, this struck me as weird. If I take the top 3 search engines that are sending traffic my way, ...
Steve Yegge at Google talks about the features of the Next Big Language. He apparantly has inside information about the language corporates are likely to make a big push for. The comments seem to suggest Javascript 2.
A timeline of Microsoft, Google and Yahoo acquisitions.
Presentations and videos from Google’s conference on test automation at London. I was at the event, and learned a lot.
Google’s internal subdomains.
Google’s master plan.
Berkeley webcasts of their courses. Google TechTalks. Authors@Google. LongNow seminars about long term thinking. UCTV Video on Demand. Nova. Computer History Museum. Comments Michelle 5 Dec 2006 12:39 pm: Do you have books I can learn From? Top 10 non-fiction books?
5 different stories about how Google was named Google. This is mentioned as the probably correct version: “Lucas Pereira: ‘You idiots, you spelled [Googol] wrong!’ But this was good, because google.com was available and googol.com was not. Now most people spell ‘Googol’ ‘Google’, so it worked out OK in the end.” Comments OverSkies 11 Dec 2008 8:29 am: Google’s founder’s son’s first words were : google.That is why they named it google
What do you find when you search Google searches? Comments Sheikh 14 Dec 2006 3:09 pm: Thank You Anand. It was very useful
55 ways to have fun with Google. Comments Dhar 25 Jun 2006 7:59 am: Just finished the book, didn’t find it too interesting. :(( S Anand 25 Jun 2006 8:52 am: Me neither. :-(
Sketchup (from Google) is now available for free download. It’s a 3D modelling tool.
Determine a sex by name. If you don’t know if a name is male or female, just search for the name on Google images. (e.g. “Priti”) Comments Umasuthan 18 Apr 2006 5:36 pm: That is not always true. Try ‘Kiran’ S Anand 18 Apr 2006 10:17 pm: You’re right. The article does warn of unisex names!
Google web authoring statistics. An analysis of over a billion pages to see how people use HTML markup.
Who’s afraid of Google? Everyone
The Blind Search Engine Test. My vote turned out to be for Google.
Google may acquire Riya. Riya can recognise faces and lettering in your pictures. Comments Sai 8 Dec 2005 11:50 pm: Hi, do post the rest of the tips for excel when you get the time. Good Stuff.
Talent wars. The interesting part was the first three paragraphs. Flying on the Delta Shuttle with Bill Gates 12 years ago, I asked, “What Microsoft competitor worries you most?” “Goldman Sachs.” I gave Gates a startled look. Was Microsoft about to try the investment banking business? “Software,” he said, “is an IQ business. Microsoft must win the IQ war, or we won’t have a future. I don’t worry about Lotus or IBM, because the smartest guys would rather come to work for Microsoft. Our competitors for IQ are investment banks such as Goldman Sachs and Morgan Stanley.” ...
Interesting experiment to test the impact of changing the rate for Google Adwords. Two identical sites have Adwords account. One has an established account paying $0.10 per word. Another is a new account, and begins by paying $1.00, then lowers it to $0.40. The former got 15,000 click-throughs regularly. The latter started just above 15,000 and fell to 1,200 after the rate drop. Does the Adwords algorithm favours rate increases disproportionately (and hence is evil?) Or is the fact that the former site an established one create the difference?
Google markets. It’s not a new Google service. Just Google “putting the wisdom of crowds to work” by using online markets.
Google Purge. Google Announces Plan To Destroy All Information It Can’t Index. Comments ritzkini 31 Aug 2005 10:28 am: hey the picture slidehsow is affecting the contact and faq section and recently updated blogs section ! S Anand 31 Aug 2005 10:57 am: Oh, what browser and version do you have? ritzkini 2 Sep 2005 12:48 pm: IE 6.0 S Anand 2 Sep 2005 3:18 pm: That’s strange… I have the same version. Having no problems. Could you mail me a screenshot?
Dodgeball lets you broadcast your location to friends via a mobile. The interesting thing is, Google has bought Dodgeball.
The authors of Freakonomics visited Google, and were asked “What would you do with our data?” BTW, there is a regular Freakonomics column on the Times.
Google supports wildcards
Taking pictures at low shutter speeds without a tripod. Comments Anonymous Hero 5 Jul 2005 9:56 am: About why Google History is important: I had searched for something from home laptop and I was at office and needed the same info. Went into Google history and found the same thing… So the search history and the links I clicked on were available from any computer attached to the Net. Found it pretty useful… Ram 5 Jul 2005 4:51 pm: Pls update the placement section with new useful links… S Anand 5 Jul 2005 6:36 pm: Ah… that’s a good use. I was also wondering if we could do some profiling of the kinds of sites we visit – perhaps by checking what their tags are on technorati or del.icio.us sathish 22 Jul 2005 11:14 am: probably based on our search history, there could some recommendations to given
Who will Google buy next? Comments Mala 22 Jun 2005 4:44 am: London pahunch gaye?? Send me your new email ID at malavikap yahoo co Anonymous 22 Jun 2005 10:23 am: anand can you give me some CAT tips tech freak 29 Jun 2005 7:03 am: give your new email ID S Anand 29 Jun 2005 8:01 am: root underscore node at yahoo dot com
Compare Yahoo and Google searches. This site lets you compare the rankings of each site on identical searches with Yahoo and Google.
Firefox prefetches Google results.
You are currently using 18 MB (2%) of your 1101 MB. Gmail storage increases by 1MB per minute. They are planning to hike capacity to 2GB. Exactly 1 year after launch.
code.google.com Comments ritzkini 24 Mar 2005 5:18 am: google rocks !!
On Google going public.
Google donates infrastructure to Wikipedia. Possible benefits to Google? Test another end use for the famed Google OS Get an “authoritative” knowledge base to provide search results on Position against Microsoft Encarta as an encyclopaedia
Google blogger fired, possibly for the blogging. Maybe he should talk to the Committee to Protect Bloggers. via MetaFilter
On the impact of ‘rel=nofollow’ tag. Or, perhaps, some prominent webloggers will use “nofollow” as a money-making technique: for a fee, they’ll take the “nofollow” off your link and let you have their PageRank.
Google Scholar lets you search academic references (journals, papers, etc).
Microsoft bought Lookout. Lookout searches for e-mail in Outlook across .pst files. As soon as they did, Microsoft closed downloads of Lookout. In contrast, when Google bought Picasa, they opened it up for free.
Picasa is good. I incorporated all my photos in it, and I can now search for photos of people by name. (I have catalogued my photos’ file names with peoples’ names.) Now, with Google taking over Picasa, we will soon be able to store/share photos on Google, and search across others’ shared photos, using Google’s image search!
I have 3 gmail invites available. Mail me if you’re interested (and WHY you’re interested in gmail).
I have no gmail invites left. Please wish Rajneesh, Joseph, Kannan, Mark, Mohamed and Naresh all the best with their accounts.
I have 6 gmail invites available. Mail me if you’re interested (and WHY you’re interested in gmail).
Google IPO. Their SEC registration is very readable – and gives a perspective on how corporate governance should be.
Bruce deserves Gmail. Context: Google has invited active bloggers on Blogger.com to use Gmail Beta. Bloggers who joined post facto have not been invited. (I can empathise with what Bruce feels.)
On the new features of search engines.
Google doesn’t think it’s the top [search engine](http://www.google.com/search?q=search engine) any more. via Anders Jacobsen
Google lets you search by location in the US. via GoogleBlog
Google is fiddling around with algorithms to make PageRank 5 times faster. Wonder why they want to do it. The article mentions something about personalised search engines, but I don’t quite get it.
About PageRank
googlemenus.blogspot.com.
An interesting Google result for “French military victories”. Unfortunately, it no longer returns the same result. But sure was an interesting find. via RHF
Google hacks on BuzzToolBox.com.
Word Spy’s top 100 words has Google on top right now, probably because Google asked them not to verbify them.
Google buys Pyra. Pyra runs Blogger and BlogSpot. via Scripting News
A blogger’s first week at Google. Ovidiu Predescu. via GoogleBlog
Very interesting article on Google on Wired. One interesting point the article raises towards the end is this: “As a private company, Google has one master: users. As a public company, there are shareholders to worry about.” And the interests of these may not be aligned. Shareholders may want more ad revenues. Users do not want ads. Shareholders may want paid placements. Users do not. Once Google IPOs, I suspect its quality will fall.
Froogle. A product search from Google. The way Google is going, I don’t think there’s any point in most sites bothering about design. Google will just offer it to customers the way they want to see it. via Kiruba
Kiruba’s interview with Google. Nice questions! And good to see the post on MetaFilter. via MetaFilter
On figuring Google’s pagerank. It’s interesting to note that Google, and only Google, ranks 11 out of 10 in its pagerank.
Now Google is complying with local laws by eliminating anti-Semetic sites from their French (google.fr) and German (google.de) sites. These sites still remain accessible through google.com, though. In the long run, I don’t think this state of affairs will continue. I hope we have one law across the world, at least where the Net is concerned. But more likely is an agreement on law guidelines, something like the WTO. Less likely and desirable is where tracking surfers’ country of origin becomes reliable. ...
Google games on Metafilter.
Get an idea of how much trouble copyright is generating for Google. via LinuxJournal
Nostalgia. Remember when Microsoft was more evil than Satan himself?
Kamat observes that most people find out about the site via a search in Google. I’d say advertising is probably pointless for just traffic generation. If one has something to sell, fine. Not otherwise.
I know – I’ve been talking too much about Google. Still, here’re the dates on which Google updates itself.
PigeonRank. Google’s amazing technology (funny).
Google answers questions.
Here’s the current status on Google vs Church of Scientology.
news.google.com. Still in Beta, and a little unimpressive, but as with most things at Google, likely to become a de facto search engine.
I searched on Altavista after a long time (oh, for no other reason than the fact that Google said I could also try my searches on Altavista, Lycos, Yahoo, etc.) and I was surprised how much the search results resembled Google’s.
The original Google paper.
Google fights googlebombing… or does it?
Google speaks Klingon, Hacker, Hindi, Tamil, and several other new languages. (Be careful about the “Save preferences” button, if you picked a language you don’t understand.)
Googlewhacking.
Google reduces need for domain names.
Life at Google still sounds like fun.
2001 in retrospect: Google.
The evolution of Google post the Sep 11 crash.
Teoma and Wisenut are search engines like google. Teoma’s “expert links” feature looks promising.
The Machiavellian Intelligence hypothesis says that the brain evolved more for its social purpose, than for finding food and things like that. Incidentally, Google’s collection on evolution is as comprehensive as ever.
Google just goes on. They’re working on voice search now.
Danny Yee’s micro-advertising experiment indicates that weblogs like Robot Wisdom (which is quite impressive) may be more effective than Google.
Google bought Deja. I don’t know which is better news – that they have money to buy companies, or that I can now search the Usenet.
Google likes directory sites. Google likes Yahoo!. Google uses peer review. Google is good.
I’ve always liked devices attached to the Internet. Web cameras are a hot favourite. Earthcam is a portal for web cameras. Google has more webcam directories.
Curious that Google should return Microsoft’s home page when you search for “more evil than Satan himself”… When I tried this on 24th June 2000, it didn’t work. Maybe Google got smart. Or was it Microsoft? On 3rd July 2000, it worked again. Comments Anonymous 18 Dec 2006 3:11 am: I tried it, doesn’t work now Max 30 Mar 2009 6:26 pm: It’s been fixed, Microsoft (and also Bush) were pretty mad at Google.