close this bookVolume 10: No. 06
View the documentInvestment
View the documentEntrepreneurship
View the documentElectronic commerce
View the documentJournal calls
View the documentAI languages
View the documentRobotics

For those US investors who got IPO stock at the offer price last year, the median appreciation over one day was 30%; over three months it was 26%. (Average increases, rather than median, were 64% and 104% due to the few really high flyers.) Median appreciation for tech IPOs was 63% in one day and 116% over three months. However, the offer price is typically available only to favored investors. If you bought in at the day's closing price, the median appreciation over three months was 0% across the board or 24% for tech stocks. There are enough people who want in on the deal that 1999 was a record year for IPOs. 555 companies went public -- almost half of them tech companies -- raising almost $74B. That's nearly 50% more than the record 1996 capitalization.

US venture funding is increasingly flowing to communications, software, and information-processing businesses, which are getting 60% of new investment (up from 41% in 1995). Health care and biotech have dropped from 20% to just 7%. VCs invested $21B in the first three quarters of 1999, 50% more than in all of 1998 and four times as much as in 1995.

US venture capital is concentrated in California and the Northeast, and in knowledge-intensive industries. (Ditto for technology advances and new patents.) 75% of 1999 venture capital has gone to CA, MA, NY, TX, WA, and CO. Most of the wealth from previous venture funding is concentrated in those states, as are skilled knowledge workers and tech-related university programs. [Gene Koretz, BW, 07Feb00, p. 30.]

Arthur Andersen Ventures is a new, $500M venture capital fund for Web startups, charging equity instead of fees. It will invest in business-to-business commerce and Internet-based services. [WSJ, 24Jan00. NewsScan.]

Softbank, the Japanese computer publishing and Internet conglomerate, is forming a $500M Softbank Emerging Markets investment fund to invest in Internet startups in about 100 developing countries. (The World Bank's International Finance Corporation is providing about 25% of the funding.) [Financial Times, 14Feb00. NewsScan.]

Internet companies raised almost $20B in venture funding last year, up from $3.4B in 1998. More than half of all 1999 venture investments went to Internet-related businesses. Funding in 2000 will likely shift to business-to-business services, rather than consumer startup such as Webvan and [PricewaterhouseCoopers. IBD, 14Feb00. NewsScan.]

When startup Direct Hit needed venture funding in 1998, Warren Packard of Draper Fisher Jurvetson arranged the investment within one day of hearing of it. (A record, to be sure.) Now Ask Jeeves Inc. (Wellesley, MA) is buying Direct Hit for $500M in stock, and DFJ will get about $105M from its $3.75M investment over three rounds of financing. That's a 28-to-one return, the kind of event that keeps Silicon Valley investment churning. VCs are often afraid *not* to invest, and their big scores keep money flowing in from elsewhere. (Then they pressure the startups to grow quickly so that the money can be recovered and put back into play.) VCs pumped almost $6B into 362 Bay Area companies in the 4th quarter (up from $1.2B in 4Qtr98), with average 1999 investment of $12M in 1,153 companies. (That's a doubling of deal size and of total investment, to $13.5B for the year. A few of the deals were over $175M; Webvan group got $275M.) Accel Partners shepherded to IPOs this past year, bringing a $5.6B return on $56M invested. It also took in 45 new companies, for $372M -- double what it could invest in 1998. One theory is that entrepreneurial demand is the "external energy source" driving this seeming perpetual motion machine (or success spiral, or snowball effect): as entrepreneurs develop the desire and need for funding, VCs will find a way to raise the capital. Stock options and capital gains tax cuts are also important drivers, as is technological progress. Growth is rapid partly because intellectual property can be accumulated much faster than physical property. Another theory says that the stock market's attitude toward Internet IPOs -- which could change at any time -- is the chief energy source.

Bay Area startups are also getting funding from corporations, buyout funds, investment banks, and financial corporations -- possibly $3B in corporate funding last year, about nine times the 1998 total. Corporations usually invest in strategic technologies -- outsourcing their R&D -- rather than just for maximum profit, so they may be willing to invest more than would a VC (though often not with the same speed). They may provide advice, administrative support, technology access, standards development, publicity, and sales channels, but their interest sometimes falls off quickly after they gain access to the technologies they want. In other cases they are investing to speed up the development of a new market. They may not be as interested as VCs in getting to an IPO quickly, so they don't take board seats or provide the same type of executive guidance. The emphasis is on product and market success -- thus helping and enriching the corporate partner -- rather than founders' profit. Often a corporate sponsor will work hand-in-hand with a VC, for the best balance of financial support. Intel invested more than $1B in 250 companies last year, at about $1M-$10M per company. (Its total portfolio is about $8B in 350 companies worldwide.) PricewaterhouseCoopers is just beginning to track these sources, with existing estimates compiled by Venture Economics (Newark, NJ).

VCs invested almost $36B in the US, up from $14B in 1998. New England got $4B; the New York City area, $2.5B. In the Bay Area, $3B went to software companies. Of the $15B in 4th-quarter VC money, the Bay Area got 39%; New England, 11%; the Southeast, LA/Orange County, and New York each 7%; DC, the Midwest, Northwest, and Colorado each 4%. 15% went to software companies (or 23% in the Bay Area). The limiting factor at this point may be people rather than funding. [Shawn Neidorf , SJM, 13Feb00, 1E and 3E.]

Part of the game is to get backing from prestigious VCs. "If Kleiner Perkins thinks enough of you to invest, or so goes the thinking, other people will surely follow the lead." It's been great for the VCs, who typically invest only pennies to perhaps $1 in stock that sells for $10 or more. Almost all of the risk has been moved to the public markets, giving extraordinary returns to all levels of pre-IPO investors. If anyone is taking stupid risks, it's the small investors. "The skinning of momentum-maddened, greedy suckers is a necessary part of capitalism." When prices finally drop, it may be by as much as 90%. After that, even the best ideas will have trouble getting funded. Most VCs are already saying that the market is nuts and that bizarre and stupid ideas are being funded -- except for their own portfolios. [Dan Gillmor, SJM, 13Feb00, 1E.]

Still interested in day trading and high-tech stock investment? Send a "subscribe" message to . [BASES, 24Jan00.]

----- "Shoot ducks when they're flying. Investors determine when investors are interested -- and that's when entrepreneurs must pull the trigger." -- Moneyhunt Rule #27. -----

It's hard to get started in Bangalore, India -- what with spreading bribes around even to get a phone line -- and most tech ventures fail. Still, it's getting better in Silicon Plateau. Many bright, well-trained Indian engineers are starting their own companies instead of working for multinationals. Mahatma Gandi Road in downtown Bangalore is crowded with shoestring startups in cubbyhole offices, sometimes with poor access to electricity or even restrooms. Often the corporate email address is a free Hotmail account. Web/software developer IQura Technologies is one of the successes, having reached 18 employees and over $300K in revenue in 14 months. (The four founders recently turned down a buyout offer of $1.4M.) They're even planning to offer stock options, though recruiting good people has not been a problem. The Indian software industry is noted for low costs, high quality, and excellent service, but seldom for entrepreneurial pizazz. Now a new generation is being inspired by Infosys and other Indian successes. As Indian companies develop more work than they can handle, employees are starting to think that they could spin off a portion for themselves. The industry now does $4B in business, up from $150M ten years ago. IQura started with just $35K and a $400 project from a founder's brother-in-law. If it hadn't worked out, they could have gone to work for HP, Siemens, TI, or a big Indian company. [Mark McDonald, SJM, 07Feb00, 1E.] (One of the founders shifted to computer science because "that's what all the girls were studying." I hope it worked out for him.)

Narayana Murthy, 53, co-founded Infosys (Bangalore) in 1981 with six friends and $1K. He used to be a socialist, but realized that you have to create wealth before it can be spread around. Now hundreds of his employees are dollar millionaires or rupee millionaires. Infosys is one of the two or three largest companies in India, across all sectors, with a capitalization of $25B. The company had 5K employees worldwide, mostly quite young (average age 26). The 42-acre campus in Bangalore offers a gym, tennis courts, and other amenities. New government policies in 1991 helped to jump-start Infosys and the Indian software industry. [Ibid.]

Latino businesses are also doing well in Silicon Valley, growing at twice the rate of other California businesses -- and Latina-owned businesses are growing even faster. Most are non-tech businesses, in an area where the cost of living is very high -- but computing hardware, software, sales, and service companies are emerging, including Internet and e-commerce businesses. The Silicon Valley Hispanic Chamber of Commerce is working for a tighter community with better information flows. One of their educational thrusts is to develop the bookkeeping practices and formal business plans that are needed to draw investment. [Alex Torres. K. Oanh Ha, SJM, 07Feb00, 1E.]

Check out the Stanford University Entrepreneurship Portal at , describing Stanford's outreach and education programs. [BASES, 24Jan00.]

----- "What makes equality such a difficult business is that we only want it with our superiors." -- Henry Becque. -----

About 97% of Webster's dictionary words are now registered for Internet use. The name has sold for $1M, for $2.2M, and for $7.5M. is one of the domain-name brokers. [AP. SJM, 14Jan00. NewsScan.]

The Open Forum Discussion List covers Web page creation, website design, e-publishing, search engines, and tips on marketing and operating a business. . [ or , NEW-LIST, 20Jan00.]

The 345-person town of Halfway, OR, has voted to rename itself to help attract new jobs. They will get about $110K in computer hardware and consulting from founders of, an Internet company selling used consumer products. [Reuters. BENT-news, 21Jan00.]

When time-stamps matter, who makes sure your systems clock is accurate? CertifiedTime will be offering a "DirectConnect" service to create a legal audit trail and evidentiary witness capability for e-commerce applications. System clocks will be locked to NIST's federally traceable timeservers. [E-Commerce Times, 26Jan00. Edupage.] (I dunno, this just sort of violates the whole spirit of the Internet...)

"14% of US consumers will visit a Web site listed in an unsolicited email message" according to an Ernst & Young survey, with something like 85% of the responses coming within 48 hours. The click-through rate is three to 10 times that of banner ads. [LA Times, 24Jan00. NewsScan.] ("Will visit" under what circumstances? Anyway, that's the reason we see so much spam.)

The Direct Marketing Association (DMA) now has an E-Mail Preference Service (e-MPS) where you can register your desire not to receive unsolicited commercial messages. . [NY Times Online, 11Jan00. Edupage.]

SpamCop claims to have a way to block or get back at spammers. . [Jo-Ann Stanton, NewsScan, 26Jan00.]

WebSiteDaily announces new web sites, resources, and events. Send an email to , or visit . List archives are at . [Robert W. Neill, Jr. , net-hap, 12May99.]

"NewsHub Filtered" summarizes daily stories from dozens of news sources, for busy Internet professionals. . [, NEW-LIST, 12Aug99.]

----- "A first-rate soup is more creative than a second-rate painting." -- Abraham Maslow. -----

Evolutionary computing and neural networks; J. of Integrated Computer-Aided Engineering, early 2001. 01Mar00; Xin Yao . . [connectionists, 26Jan00.]

Humanoids; IEEE Intelligent Systems, Jul00. 07Mar00 (with abstracts already due); Mark L. Swinson , (703) 696-2247, (703) 696-0564 fax. . [Bill Park, 28Jan00.]

Metadata; Interactive Learning Environments (ILE). 14Apr00; , +32 16 327066, +32 16 327996 fax. or . [dbworld, 18Feb00.]

Artificial intelligence in environmental engineering; Computer-Aided Civil and Infrastructure Engineering (CACAIE). 01May00; Miquel Sanchez-Marre . . [, 18Feb00.]

----- "What is Art? It is the response of man's creative soul to the call of the Real." -- Rabindranath Tagore. -----

Brian T. Rice is seeking people interesting in working on Slate, a new programming language combining dynamic elements of Self with the pattern-programming and abstractions of BETA, plus Lisp's abstract syntax tree, Smalltalk-80's recursive interpreter/compiler/interface, and a functional object semantics. (One interface project is a language and environment that is programmable by visual gestures and representations alone.) Slate's code will be released under GPL. A partial prototype has been implemented under Common Lisp. . [ or , comp.lang.lisp, 12Jan00.]

The CLISP expert system shell project has moved to . Source and binaries can be downloaded from or , with snapshots and diffs available at . CLISP mailing list subscribers must re-enroll in clisp-announce or the clisp-list discussion list (or clisp-devel maintainers' list) at . [Sam Steingold , comp.lang.lisp, 14Jan00.]

Digitool has recently released its MCL 4.3 (PPC) and MCL 3.4 (68k) Lisp environments for the Mac. . [Rainer Joswig , comp.lang.lisp, 15Jan00.]

Poplog is a multi-language environment developed at USussex. The core is POP-11, which is essentially a Lisp with a front-end parser:

define sum(list); if null(list) then 0 else hd(list) + sum(tl(list)) endif enddefine;

POP-11 has a built-in incremental code-generator for a virtual machine, supporting free implementations of Common Lisp, SML, Prolog, and Scheme. There is also an X-windows environment, with an editor and the ability to write X-applications. Host systems include many workstations, Linux, and Windows. , or for UMASS Scheme. [Robin Popplestone , comp.lang.scheme, 01Feb00.]

XLisp is another Mac Lisp, still going strong as part of Xlisp-Stat. . [Simon P. Blomberg , comp.lang.lisp, 17Jan00.]

For Lisp info and book recommendations, see and . [Andrew Cooke and Michael Hudson , comp.lang.lisp, 04Feb00.]

----- "Language is the armory of the human mind, and at once contains the trophies of its past and the weapons of its future conquests." -- Samuel Coleridge. -----

You can find info about Robot Wars (including two 1998 BBC programs), BattleBots, Survival Research Laboratories, SEEMEN robotic/kinetic art, Machinus/Artifacts of Mass Distraction, etc., at . [Bill Park , 15Jan00.] (Apparently the Robot Wars program is one of the most popular on British TV. Stations in the US are starting to take an interest.)

The Logo Foundation is a nonprofit organization supporting use of Logo-based software and learning environments. . [Amnon Till , net-hap, 20Sep99.]

Robotics and helicopters is the subject of Mike Fouche's website at . [,, 14Dec99.]

Personal Robotics News is of interest to those building their own robots. You can read current and back issues at ("Mailing List"), or sign up for a free subscription. You can also get free plans for the FirstBOT Mobile Robot at , or a kit at . [Don Roy ,, 07Feb00.]

A new adaptive robotics website offers paper reviews, conference reports, links, and tutorials at . [Bill Park , 28Jan00.]

Getting robots to work together is difficult. You can see pictures and videos of robot swarms and work groups from last year's Khepera Workshop, at . [Francesco Mondada , sci.nanotech, 09Feb00. Bill Park.]

-- Ken